Agribusiness Management · Foundations · also ABM 541, Computer Applications for Agribusiness, MIS (NIAM)
Working computer knowledge for a manager, from spreadsheets and statistics packages to databases, cloud, cyber-security and management information systems.
In the exams: Computer knowledge sections of agricultural development bank officer recruitment (qualifying); light in agricultural master's entrance examinations
Computer applications and management information systems (MIS) is the working computer knowledge a manager needs: how the machine and its software are built, how data are stored and queried, how spreadsheets and statistical packages turn records into decisions, how networks, the web and the cloud carry those data, how to keep them safe, and how an organisation designs information systems that support each level of management. It is not a programming course: the aim is that a manager can specify a system, build a sound spreadsheet model, read a query and recognise a security risk.
The subject matters in agribusiness because the sector runs on many small transactions spread over a wide area. A dairy cooperative records milk volume and fat from thousands of members twice a day; a rural bank appraises crop loans against land records and scales of finance. Each is an information system whose quality decides whether farmers are paid on time and fraud is caught.
In India it appears as a qualifying Computer Knowledge section in the NABARD Grade A Phase I paper and similar bank and development-finance officer examinations, where questions test hardware, software, operating systems, networks, the internet, databases, office tools and security. In the ICAR AIEEA-PG Agri-Business Management paper it carries light weight. This manual follows the ten units of the curriculum, adds the systems-development and agri-information-system material that courses teach under MIS, and leaves predictive analytics, machine learning, GIS and ERP implementation detail to the companion subject Business Analytics and Digital Agribusiness.
A computer is an electronic machine that accepts data (input), processes it under the control of a stored program, stores data and results, and produces output. The stored-program idea, usually credited to John von Neumann's 1945 report, means that instructions and data sit in the same memory, so a computer can be reprogrammed without rewiring. The classic model has four parts: input devices, a central processing unit (CPU), memory and storage, and output devices, joined by buses that carry data, addresses and control signals.
Textbooks group computers into generations by the switching technology used. The first generation (roughly the 1940s and 1950s) used vacuum tubes; the second used transistors; the third used integrated circuits; the fourth, from the early 1970s, uses microprocessors built by very-large-scale integration, which put a whole CPU on one chip. 'Fifth generation' is used loosely for machines aimed at artificial intelligence and massively parallel processing. By size and role, computers range from supercomputers (weather and climate models) and mainframes (bulk bank transactions) to servers, personal computers, and embedded devices inside equipment such as milk analysers and weighbridges.
The CPU has an arithmetic logic unit (ALU), which does arithmetic and comparisons; a control unit (CU), which fetches each instruction from memory, decodes it and directs the other parts; and registers, very small and very fast storage locations inside the processor. The CPU repeats the fetch-decode-execute cycle billions of times a second; clock speed is measured in gigahertz, but cores and cache matter as much.
Memory is organised as a hierarchy from fast, small and costly to slow, large and cheap: registers, cache, main memory (RAM), then secondary storage (solid-state drives and hard disks), then removable and archival media (optical discs, magnetic tape, cloud archives). Random-access memory is volatile: its contents are lost when power is removed. Read-only memory (ROM) and its programmable forms are non-volatile and hold firmware such as the start-up code. Secondary storage is non-volatile and holds the operating system, programs and files.
Data are represented in binary. A bit is a single 0 or 1; a byte is 8 bits; a nibble is 4 bits. Storage prefixes come in two conventions that exams and vendors mix: the decimal (SI) convention, in which a kilobyte is 1,000 bytes and a gigabyte 10^9 bytes, and the binary (IEC) convention, in which a kibibyte (KiB) is 1,024 bytes and a gibibyte (GiB) is 2^30 bytes. Many older textbooks write KB for 1,024 bytes. Network speeds are quoted in bits per second, file sizes in bytes, so a file of 600 megabytes sent over a 40 megabit-per-second link needs about 600 × 8 / 40 = 120 seconds, not 15.
Besides the keyboard, mouse, touch screen, scanner and camera, input devices for bulk data capture include magnetic ink character recognition (MICR), used by banks to read the codes printed on cheques; optical character recognition (OCR), which converts printed text images into text; optical mark recognition (OMR), which reads filled bubbles on answer sheets and survey forms; and barcode and QR-code readers and radio-frequency identification (RFID) readers, used for stock, traceability and livestock tags.
Software is the set of programs that make hardware useful. System software runs the machine: the operating system, device drivers, and utilities such as backup, compression and antivirus tools. Application software does users' work: word processors, spreadsheets, databases, accounting packages, statistical packages and custom business applications. Programs are written in programming languages. Machine language is binary instructions; assembly language uses short mnemonics and is converted by an assembler; high-level languages are converted by a compiler, which translates the whole program into machine code before it runs, or by an interpreter, which translates and executes it statement by statement. Proprietary software is licensed under a vendor's restrictions; open-source software publishes its source code under a licence that allows change and redistribution; freeware is free of charge but not necessarily open. Licence terms decide total cost and whether the organisation can move its data elsewhere.
Computers store everything in binary because an electronic switch has two reliable states. A number system has a base, or radix: decimal uses base 10 with digits 0 to 9; binary uses base 2 with digits 0 and 1; octal uses base 8 with digits 0 to 7; hexadecimal uses base 16 with digits 0 to 9 and A to F, where A is 10 and F is 15. Each position is worth the base raised to a power, counting from zero at the right. The binary number 10110110 is 128 + 32 + 16 + 4 + 2 = 182 in decimal.
To convert a decimal integer to binary, divide repeatedly by 2 and read the remainders from the last to the first: 45 gives remainders 1, 0, 1, 1, 0, 1 from the first division to the last, so 45 is 101101 in binary. Hexadecimal is a shorthand for binary because one hex digit stands for exactly four bits: 1011 0110 is B6 in hex. Octal groups three bits.
With n bits a computer can represent 2^n distinct values. Eight bits give 256 values (0 to 255 unsigned); 32 bits give about 4.29 billion values, which is why a 32-bit memory address can reach at most 4 GiB of memory and why the 32-bit version of the internet protocol (IPv4) has about 4.3 billion addresses. Negative integers are usually stored in two's complement form, and real numbers in floating-point form, which stores a sign, a significand and an exponent and so cannot represent most decimal fractions exactly. That is why a spreadsheet occasionally shows 0.30000000000000004 where 0.3 was expected.
Text is stored by giving each character a number. ASCII, the American Standard Code for Information Interchange, uses 7 bits and so defines 128 characters, enough for English letters, digits and control codes. Unicode assigns a code point to characters of almost every writing system, including the Indian scripts, Chinese, Arabic and Amharic; UTF-8 is the most common way of storing Unicode, using one byte for the ASCII range and up to four bytes for other characters. A farmer database that must hold names in local scripts must use Unicode throughout, or names will be corrupted when data move between systems.
A raster image is a grid of pixels, each stored as a colour value. An uncompressed image of 1,920 by 1,080 pixels with 24-bit colour (3 bytes per pixel) needs 1,920 × 1,080 × 3 = 6,220,800 bytes, about 5.9 MiB. Compression reduces this: lossless compression keeps every bit and is used for text, program files and archives; lossy compression discards detail the eye or ear is unlikely to notice and is used for photographs, audio and video. Vector graphics store shapes as geometry and scale without loss.
Comma-separated values (CSV), a plain-text table format, can be read by almost any program and is the safest way to move tables between a spreadsheet, a database and a statistical package, though it carries no formulas, formatting or data types. A system contract should state the open formats in which the organisation can export all of its data.
An operating system (OS) is the system software that manages the computer's hardware and provides services to application programs and users. Its core, the kernel, stays in memory and handles the most basic tasks; the shell or user interface lets people give commands, either by typing them in a command-line interface or through a graphical user interface (GUI) of windows, icons, menus and a pointer.
The OS has five main management functions. Process management schedules processes (programs in execution) on the CPU. Memory management allocates RAM to processes and protects one process's memory from another; with virtual memory it divides memory into fixed-size pages and moves pages between RAM and disk as needed. If the system spends more time swapping pages than doing useful work, it is said to be thrashing, and the cure is more RAM or fewer simultaneous programs. File management organises data into files and directories (folders), controls access permissions and keeps track of free space. Device management runs input and output through drivers; security functions authenticate users and enforce permissions.
When a computer is switched on, firmware (the older BIOS or the newer UEFI) runs a power-on self-test (POST), then finds and starts a boot loader, which loads the OS kernel into memory. A cold boot starts from power-off; a warm boot restarts without cutting power.
Operating systems are classified by how they share the machine. A batch system runs jobs one after another without user interaction, as early payroll systems did. Multiprogramming keeps several programs in memory so that the CPU switches to another when one waits for input or output. Time-sharing, or multitasking, gives each user or program a short slice of CPU time in turn, so that many users feel they have the machine to themselves. Multiprocessing uses two or more CPUs or cores. A real-time OS guarantees a response within a fixed time and is used in control equipment such as irrigation controllers, milking robots and medical devices. A distributed OS manages a group of networked computers as one system.
Scheduling decides which ready process gets the CPU next. First-come, first-served (FCFS) runs jobs in arrival order; it is simple but a long job makes short jobs wait. Shortest job first (SJF) runs the shortest job next and gives the lowest average waiting time when all jobs are known, but long jobs may starve. Round robin gives each job a fixed time quantum in turn and suits time-sharing. Consider three jobs arriving together with CPU bursts of 10, 4 and 2 milliseconds in that order. Under FCFS their waiting times are 0, 10 and 14, an average of 8 ms. Under SJF the order is 2, 4, 10, the waiting times are 0, 2 and 6, and the average is about 2.67 ms. Under round robin with a 4 ms quantum, the jobs finish at 16, 8 and 10 ms, so their waiting times are 6, 4 and 8 and the average is 6 ms.
A deadlock occurs when processes each hold a resource and wait for one held by another. Four conditions must hold together: mutual exclusion, hold and wait, no pre-emption and circular wait. Removing any one prevents deadlock.
A database is an organised collection of related data; a database management system (DBMS) is the software that stores, retrieves, secures and updates it. Before DBMSs, each department kept its own files, which caused data redundancy (the same farmer's address in several files), inconsistency (the addresses differ), difficulty in answering new questions, weak security and problems when two people updated the same record. A DBMS solves these by holding data once, under central control, with a data dictionary (metadata describing every table and field), access control, backup and recovery, and concurrency control.
The ANSI/SPARC three-schema architecture separates the external level (each user's view), the conceptual level (the logical structure of the whole database) and the internal level (physical storage). This gives data independence: physical storage can change without changing programs (physical independence), and the logical structure can be extended without changing every user view (logical independence).
Older hierarchical (tree) and network (graph) models gave way to the relational model, proposed by E. F. Codd in 1970. In the relational model data sit in tables called relations. Each row (tuple) is a record; each column (attribute) holds values from one domain. A super key is any set of columns that identifies a row uniquely; a candidate key is a minimal super key; the primary key is the candidate key chosen as the main identifier and may not be null; a composite key has more than one column; a foreign key is a column that refers to the primary key of another table and so links the two. Entity integrity requires primary keys to be non-null and unique; referential integrity requires every foreign-key value to match an existing primary key or be null. Non-relational (NoSQL) databases trade some of these guarantees for flexibility or scale.
Database design starts with an entity-relationship (ER) model. An entity is a thing about which data are kept (farmer, plot, loan, delivery); an attribute is a property of an entity (name, area, amount); a relationship links entities (a farmer owns plots). Cardinality states how many: one-to-one, one-to-many (one farmer, many deliveries) or many-to-many (many farmers sell to many traders), which in a relational design needs a linking table.
Normalisation removes redundancy and update anomalies by splitting tables according to functional dependencies. A table is in first normal form (1NF) when every field holds one atomic value and there are no repeating groups, so 'crop1, crop2, crop3' columns or a list of crops in one cell violate it. Second normal form (2NF) requires 1NF and that no non-key attribute depends on only part of a composite key. Third normal form (3NF) requires 2NF and that no non-key attribute depends on another non-key attribute (no transitive dependency): if a delivery table holds village and also the village's district, district depends on village rather than on the delivery, so it belongs in a village table. Boyce-Codd normal form (BCNF) is a stricter form of 3NF in which every determinant is a candidate key.
A transaction is a unit of work that must be completed in full or not at all, such as debiting one account and crediting another. Reliable DBMSs guarantee the ACID properties: atomicity (all or nothing), consistency (the database moves from one valid state to another), isolation (concurrent transactions do not see each other's half-finished work) and durability (a committed change survives a crash).
Structured Query Language (SQL) is the standard language of relational databases, standardised by ANSI and ISO; vendors add their own extensions, so details vary. Its commands fall into groups. Data definition language (DDL) defines structure: CREATE, ALTER, DROP and TRUNCATE. Data manipulation language (DML) works on rows: SELECT, INSERT, UPDATE and DELETE. Data control language (DCL) manages permissions: GRANT and REVOKE. Transaction control language (TCL) manages transactions: COMMIT, ROLLBACK and SAVEPOINT. Some textbooks place SELECT in a separate data query language.
A basic query reads: SELECT columns FROM table WHERE row-condition GROUP BY columns HAVING group-condition ORDER BY columns. The logical order of evaluation is FROM, WHERE, GROUP BY, HAVING, SELECT, ORDER BY. WHERE filters individual rows before grouping; HAVING filters groups after aggregation, so 'villages whose total delivery exceeds 5,000 litres' needs HAVING SUM(litres) > 5000. Aggregate functions are COUNT, SUM, AVG, MIN and MAX; COUNT(*) counts rows while COUNT(column) ignores nulls. DISTINCT removes duplicate rows from the result. Pattern matching uses LIKE with % for any string and _ for one character. A null means 'unknown', so it is tested with IS NULL, not with = NULL.
Joins combine tables. An inner join returns only rows that match in both tables; a left outer join returns every row of the left table, with nulls where the right table has no match, which is how one finds farmers who made no delivery this month; a cross join returns every combination, so a table of 120 farmers crossed with a table of 5 crops gives 600 rows. DELETE removes rows and can be rolled back inside a transaction; TRUNCATE removes all rows quickly; DROP removes the table itself. An index speeds up searches on a column at the cost of extra storage and slower inserts. A view is a stored query presented as a virtual table.
A spreadsheet is a grid of cells addressed by column letter and row number (B4). A cell holds a value, a label or a formula beginning with '='. When a formula is copied, a relative reference (B4) shifts with it, an absolute reference ($B$4) stays fixed, and a mixed reference ($B4 or B$4) fixes only the column or only the row. Copying =B2*$C$1 from D2 to E5 gives =C5*$C$1, because the relative part moves one column right and three rows down while the absolute part stays.
The functions every manager uses fall into a few groups, and their names and argument order are broadly similar across spreadsheet programs, though details vary. Mathematical and statistical: SUM, AVERAGE, MIN, MAX, COUNT (numbers only), COUNTA (non-empty cells), ROUND, and the conditional forms SUMIF, COUNTIF and AVERAGEIF with their multi-condition versions. Logical: IF(test, value-if-true, value-if-false), AND, OR, and nested IFs for slabs. Lookup: VLOOKUP(value, table, column number, match type) searches the first column of a table; with exact match it returns an error if the value is absent, and with approximate match it needs the first column sorted ascending and returns the row with the largest value not greater than the one sought, which is exactly how commission or tax slabs are applied. INDEX with MATCH does the same job without the restriction that the lookup column be on the left. Text functions (LEFT, MID, TRIM) clean imported data; financial functions include PMT, PV, FV, NPV and IRR.
Two financial functions catch students. PMT returns the level instalment on a loan; for a loan of 500,000 at 9 per cent a year repaid in five annual instalments, PMT gives about 128,546 per year, and with monthly instalments the rate must be divided by 12 and the periods multiplied by 12, giving about 10,379 a month. The spreadsheet NPV function assumes the first cash flow comes one period from now; so a project costing 100,000 today and returning 30,000 a year for five years at 10 per cent has NPV = -100,000 + NPV(10%, five flows of 30,000) ≈ 13,724, whereas putting the initial outlay inside the function discounts it one year too many and gives about 12,476.
Common error values tell the user what went wrong: #DIV/0! for division by zero, #N/A when a lookup finds no match, #REF! when a formula refers to a deleted cell, #VALUE! when text sits where a number is expected and #NAME? when a function name is misspelt. A circular reference is a formula that depends on its own result.
A pivot table summarises a list of records by grouping on one or more fields and aggregating another: for example, total litres by village and month from a year of daily collection records. Rows, columns, values (sum, count, average) and filters are dragged into place, and the table can be regrouped in seconds. A pivot table works only on a clean list: one header row, one record per row, no blank rows or merged cells, and consistent spellings, since 'Rampur' and 'Rampur ' are two villages to the software. Pivot tables must be refreshed after the source data change.
What-if analysis asks how outputs respond to inputs. Goal seek finds the one input value that makes a formula reach a target: if fixed costs are 240,000, variable cost is 18 per kilogram and expected volume is 20,000 kilograms, goal seek on profit = 0 returns a break-even price of 30 per kilogram. A one- or two-variable data table recalculates a result for a list of input values, such as profit at several prices and yields. Scenario tools store named sets of inputs (pessimistic, base, optimistic) and report the outputs side by side. An optimisation add-in solves linear and non-linear programmes, such as least-cost feed formulation or a crop plan under land, water and labour limits.
Field studies have repeatedly found errors in a large share of operational spreadsheets, so discipline matters. Separate inputs, calculations and outputs, preferably on different sheets; put each assumption in one labelled cell; never type a number into a formula; use data validation to restrict entries to allowed values; check totals by a second route; and lock formula cells. For large, multi-user or audited records, a database is the right tool.
The spreadsheet sits in an office suite with a word processor and presentation software. The word processor's mail merge combines a template letter with a data source, such as a member list, to produce one personalised letter or label per record; styles generate a table of contents; track changes supports review.
A statistical package is software built for managing data and running statistical procedures: descriptive statistics, cross-tabulations, tests of significance, analysis of variance, regression, factor and cluster analysis, and survival and time-series models. Agribusiness and agricultural-economics courses commonly teach one menu-driven commercial package, one command-driven commercial package with a separate data-step and procedure-step structure, and increasingly an open-source statistical programming language. Compared with spreadsheets, packages handle large data sets, missing values and complex survey designs better and keep a log of every step.
Packages share a common structure. A data editor shows cases in rows and variables in columns, like a spreadsheet, but each variable also has metadata: a name, a type (numeric, string, date), a label, value labels (1 = male, 2 = female; 1 = owns irrigation, 0 = does not), missing-value codes and a level of measurement. The level of measurement (nominal, ordinal, interval or ratio) decides which statistics make sense: a mean of district codes is meaningless; a mean yield is fine.
In a command-driven package the analyst writes a program with two kinds of step: a data step reads raw data, creates and recodes variables and merges files, and a procedure step runs an analysis on the prepared data and prints results. Menu-driven packages generate the equivalent command syntax behind every dialogue box, and good practice is to paste and save that syntax.
Most agribusiness data analysed in packages come from surveys of farmers, consumers, traders or firms. The workflow is: design the questionnaire with a codebook that fixes variable names, codes and missing-value rules; enter data with validation; check for out-of-range values, duplicates and inconsistent answers; recode and compute derived variables; analyse; and report. Missing values must be declared. If a survey codes 'no answer' as 999 and the analyst forgets to declare it, five yields of 12, 15, 999, 18 and 15 quintals give a mean of 211.8 instead of the correct 15.0 from the four valid answers. Declared missing values are excluded from calculations, and the analyst must also decide between listwise deletion (dropping any case with a missing value in any variable used) and pairwise deletion (using all available pairs for each statistic), which can give different sample sizes in one table.
The procedures in a typical MBA project follow the research question. Frequencies and descriptives profile the sample. Cross-tabulation with a chi-square test checks association between two categorical variables; the degrees of freedom are (rows - 1) × (columns - 1), so a 3 × 4 table has 6. Independent-samples t-tests compare two group means; one-way ANOVA compares three or more; correlation and linear regression model relationships; logistic regression models a yes-or-no outcome such as adoption of a technology. The theory belongs to business statistics; this unit is about running procedures correctly and reading output.
Reproducibility is the main professional standard. Keep the raw data untouched and do all cleaning in a saved, commented script, so anyone with the raw file and script obtains the same tables.
A computer network links devices so that they can share data and resources. Networks are classed by span: a personal area network (PAN) connects devices around one person; a local area network (LAN) covers a building or campus; a metropolitan area network (MAN) covers a city; a wide area network (WAN) spans regions or countries, and the internet is the largest WAN. Wireless LANs follow the IEEE 802.11 family of standards; wired LANs mostly follow the Ethernet (IEEE 802.3) standards.
Topology is the layout of connections. In a bus topology all devices share one cable, so a cable fault stops the network; in a star, each device connects to a central switch, so one failed cable affects one device but the central device is a single point of failure; in a ring, data pass from device to device around a loop; in a mesh, devices have multiple paths, which gives resilience at higher cost; a tree is a hierarchy of stars. Network devices include the hub (repeats signals to all ports), the switch (forwards frames only to the port of the destination device, using hardware MAC addresses), the router (forwards packets between different networks using IP addresses), the modem (modulates and demodulates signals over telephone, cable or radio lines), the gateway (connects networks that use different protocols) and the wireless access point. Optical fibre offers the highest capacity and immunity to electrical interference.
The OSI reference model divides network communication into seven layers: physical, data link, network, transport, session, presentation and application. The internet itself uses the simpler TCP/IP model with four layers: link, internet, transport and application. Bandwidth is the capacity of a link; latency is the delay.
The internet grew from ARPANET, a United States research network first connected in 1969, and is a global network of networks that share the TCP/IP protocol suite. The Internet Protocol (IP) addresses and routes packets; the Transmission Control Protocol (TCP) gives reliable, ordered delivery; UDP is faster but unreliable. IPv4 addresses are 32 bits long, written as four decimal numbers such as 192.168.1.10, giving about 4.3 billion addresses, which proved too few; IPv6 addresses are 128 bits long. In a network written as /24, 24 bits name the network and 8 bits are left for hosts, giving 256 addresses of which 254 can be given to devices, because the all-zeros and all-ones addresses are reserved. The Domain Name System (DNS) translates names into IP addresses.
The World Wide Web, proposed by Tim Berners-Lee at CERN in 1989, is a service that runs on the internet: pages written in HTML, identified by uniform resource locators (URLs) and transferred by the Hypertext Transfer Protocol (HTTP), or its encrypted form HTTPS. The internet and the web are therefore not the same thing; email, file transfer and voice calls use the internet without the web. Email is sent with SMTP and retrieved with POP3, which typically downloads messages to one device, or IMAP, which keeps them on the server and synchronises several devices. FTP and its secure variants transfer files. Cookies are small files a website stores in the browser to remember a session or track behaviour.
An intranet is a private network inside an organisation that uses internet technologies (web pages, email, TCP/IP) but is open only to its staff, for example for circulars, forms, policies and internal applications. An extranet extends part of an intranet to selected outsiders, such as dealers, suppliers, member cooperatives or a bank's correspondents, under controlled login. A virtual private network (VPN) creates an encrypted tunnel across the public internet so that a branch or travelling officer can use the intranet securely.
Electronic commerce is buying and selling over networks. It is classified by the parties: business-to-business (B2B), such as an input company's dealer ordering portal or a processor buying from aggregators; business-to-consumer (B2C), such as an online grocery; consumer-to-consumer (C2C), such as classified listings for used tractors or livestock; business-to-government (B2G) and government-to-citizen (G2C), such as online applications for subsidies. Electronic data interchange (EDI) is the older standard-format exchange of business documents between firms' computers.
In agriculture, electronic trading platforms link regulated wholesale markets so that buyers in one market can bid for produce assayed and stored in another; India's national electronic agriculture market (e-NAM), launched in 2016, is a government example, and many countries run electronic warehouse-receipt and commodity-exchange systems. These platforms depend on quality assaying, trusted weighing, warehouse receipts and timely payment as much as on software. Digital payment systems, including card networks, mobile money and real-time interbank payment systems, have cut the cost of paying many small suppliers; payment design, dispute handling and consumer protection rules vary by country and are set by each central bank or payments regulator.
HyperText Markup Language (HTML) is the markup language of the web. It is not a programming language: it describes the structure of a document, while Cascading Style Sheets (CSS) describe presentation (colours, fonts, layout) and JavaScript adds behaviour. The specification is maintained as a 'living standard' by the WHATWG; 'HTML5' is the common name for the modern form.
An HTML document is made of elements, usually written as a start tag, content and an end tag: <p>Milk procurement rose this month.</p>. Tags are enclosed in angle brackets and are not case-sensitive in HTML, though lower case is the convention. Some elements are empty (void) and have no end tag, such as <br> for a line break, <hr> for a thematic break, <img> for an image and <input> for a form field. Attributes give extra information inside the start tag as name="value" pairs, for example <a href="prices.html">Today's prices</a>.
A minimal page begins with <!DOCTYPE html>, which tells the browser to use standards mode, followed by the <html> root element containing a <head> and a <body>. The head holds information about the page that is not displayed in the page itself: the <title> shown in the browser tab and in search results, <meta> elements such as the character encoding (UTF-8) and the viewport setting for mobile screens, and links to style sheets. The body holds the visible content.
Headings run from <h1>, the most important, to <h6>. Paragraphs use <p>; emphasis uses <em> and strong importance <strong>. Lists are unordered (<ul>, shown with bullets) or ordered (<ol>, numbered), each item in <li>. A hyperlink is an anchor element, <a href="URL">link text</a>; the href may be an absolute URL with protocol and domain, or a relative path to a page on the same site; adding target="_blank" opens a new tab. An image uses <img src="file.jpg" alt="description">; the alt text is read aloud by screen readers and shown if the image fails to load, so it is required for accessibility.
Tables use <table>, with each row in <tr>, header cells in <th> and data cells in <td>; a table with three <tr> rows each holding four cells shows 12 cells. Forms collect input: the <form> element's action attribute names where the data are sent and its method is usually GET (data appended to the URL, suitable for searches) or POST (data in the request body, suitable for submissions and anything sensitive). Inside a form, <input> takes a type such as text, number, date, email, password, checkbox, radio or submit; <select> with <option> gives a drop-down list; <textarea> gives a multi-line box; <label> ties a caption to its field. Semantic elements such as <header>, <nav>, <main> and <footer> describe the role of each part of the page.
A manager seldom writes web pages but should be able to read simple markup, brief a developer, and insist that a supplier portal works on a cheap phone with a slow connection and is accessible to users with low vision or low literacy.
Cloud computing is defined by the United States National Institute of Standards and Technology (NIST, Special Publication 800-145, 2011) as a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service-provider interaction. NIST lists five essential characteristics: on-demand self-service, broad network access, resource pooling (many customers share the provider's resources, a multi-tenant model), rapid elasticity (capacity scales up and down quickly) and measured service (use is metered, which allows pay-per-use billing).
NIST names three service models. In infrastructure as a service (IaaS) the provider supplies virtual machines, storage and networks, and the customer manages the operating system, middleware and applications. In platform as a service (PaaS) the provider also manages the operating system and runtime, and the customer deploys its own applications and data. In software as a service (SaaS) the customer simply uses a finished application through a browser or app, such as hosted email, accounting, customer-relationship or farm-management software; the provider manages everything beneath. Under every model the customer remains responsible for its own data, user accounts and access settings; this division is called the shared-responsibility model, and misconfigured access settings, not provider failures, cause many cloud data leaks. Underneath is virtualisation: a hypervisor runs several isolated virtual machines on one physical server.
NIST's four deployment models are private cloud (operated for one organisation, on or off its premises), community cloud (shared by organisations with common concerns, such as a group of cooperatives or research institutes), public cloud (open to the general public and owned by a provider) and hybrid cloud (two or more of these bound together so that data and applications can move between them, for instance keeping sensitive member data in a private cloud and using public capacity for seasonal peaks).
The business case for cloud is usually framed as converting capital expenditure into operating expenditure, paying only for what is used, scaling quickly and avoiding the need for in-house server specialists. A simple comparison over four years: an on-premises server costing 1,200,000 to buy plus 150,000 a year to run costs 1,800,000; a cloud service at 30,000 a month costs 1,440,000. Comparisons must include staff time, power, data-transfer charges and the exit cost of moving data out; steady loads can be cheaper to own, spiky loads usually cheaper in the cloud.
Risks include dependence on internet connectivity (serious in rural areas, where offline-capable applications that synchronise later are often essential), vendor lock-in through proprietary formats and services, data-protection and data-residency rules that may require certain data to stay in the country (rules differ by country and change, so check the current law), and outages. Service-level agreements (SLAs) state availability as a percentage: 99.9 per cent allows about 8.8 hours of downtime a year and 99.99 per cent about 53 minutes. Edge computing processes data near the source, such as a farm gateway that summarises soil-moisture or cold-chain sensor (Internet of Things) readings before sending them, which saves bandwidth and survives connection gaps.
Information security protects three properties, the CIA triad: confidentiality (only authorised people see the data), integrity (data are accurate and not altered without authority) and availability (systems and data are there when needed). Authentication confirms who a user is, authorisation decides what that user may do, and accounting or auditing records what was done. Authentication factors are something you know (a password or PIN), something you have (a phone, card or token) and something you are (a fingerprint or face). Multi-factor authentication combines factors of different kinds; a password plus a security question is still a single factor.
Malicious software (malware) takes several forms. A virus attaches itself to a host file or program and spreads when that is run; a worm is a stand-alone program that replicates itself across networks without user action; a Trojan horse pretends to be useful software while doing harm; ransomware encrypts files and demands payment for the key; spyware and keyloggers secretly record activity and keystrokes; adware pushes advertisements; a rootkit hides deep in the system to conceal an intruder; a botnet is a network of infected machines controlled remotely, often used for distributed denial-of-service (DDoS) attacks that flood a service with traffic to make it unavailable.
Social engineering manipulates people into giving access. Phishing uses fake emails or websites to steal credentials; spear phishing targets a named person; whaling targets senior executives; vishing uses voice calls and smishing uses text messages, and both are common against rural bank customers, for example a caller posing as a bank officer asking for a one-time password. SQL injection types malicious database commands into web forms and is prevented by parameterised queries; credential stuffing reuses passwords leaked from another site. Password strength grows with length and character set: eight characters drawn from 62 letters and digits give 62^8, about 2.2 × 10^14 combinations, roughly a thousand times more than eight lower-case letters alone, yet at ten billion guesses a second that space falls in about six hours, so length, unique passwords and multi-factor authentication matter most.
Encryption turns plaintext into ciphertext using a key. In symmetric encryption, such as the Advanced Encryption Standard (AES), the same secret key encrypts and decrypts; it is fast but the key must be shared securely. In asymmetric or public-key encryption, such as RSA, each party has a public key that anyone may know and a private key kept secret: data encrypted with a recipient's public key can be decrypted only with the recipient's private key. A digital signature is made by the sender with the sender's private key over a hash of the message, and anyone can verify it with the sender's public key; it proves origin and integrity and supports non-repudiation. A hash function such as SHA-256 maps data of any length to a fixed-length digest, and any change in the data changes the digest, so hashes check integrity and store passwords safely when combined with a random salt. Digital certificates issued by certificate authorities bind public keys to identities; this public-key infrastructure underlies HTTPS. In India the Information Technology Act, 2000 gives legal recognition to electronic records and digital signatures; other countries have their own e-signature laws.
Controls are layered, so that one failure does not expose everything (defence in depth). Technical controls include firewalls that filter traffic by rules, intrusion detection and prevention systems, antivirus and endpoint protection, encryption of devices and connections, timely patching, and the principle of least privilege (each user gets only the access the job needs). Administrative controls include security policies, separation of duties, user training, access reviews and incident-response plans. Backups follow the 3-2-1 rule: three copies, on two different media, with one copy off-site or offline, which is the main defence against ransomware. ISO/IEC 27001 specifies requirements for an information security management system and is widely used for certification.
Information ethics covers privacy, accuracy, property and access. Farmer registries, loan files and mobile-money records hold sensitive data about land, income and debt. Data-protection laws commonly require a lawful basis or consent for processing, purpose limitation, data minimisation, accuracy, storage limits, security and rights for individuals to access and correct their data. The European Union's General Data Protection Regulation has applied since 2018; India enacted the Digital Personal Data Protection Act in 2023; many other countries have their own laws, and the details and enforcement differ, so a manager must check the law that applies.
Intellectual property in software is protected by copyright and licences; installing unlicensed copies (software piracy) is illegal and exposes the firm to malware. The digital divide, the gap in access and skills between rich and poor, urban and rural, men and women, is itself an ethical issue for agribusinesses that move services online: a system that only works for smartphone users with good connectivity excludes many smallholders, and a responsible design keeps an assisted or offline channel.
Data are raw facts (a delivery of 12 litres at 4.1 per cent fat); information is data processed into a form that is meaningful and useful for a decision (this centre's average fat has fallen for three weeks). Useful information is accurate, timely, relevant, complete enough, concise, and presented to the person who can act on it. 'Garbage in, garbage out'.
An information system collects, processes, stores and distributes information to support decision-making, coordination and control. A management information system in the broad sense is the whole of an organisation's information systems serving management; in the narrow sense it is the system that produces routine reports for middle managers from transaction data.
Robert Anthony's framework (1965) distinguishes three levels of management activity: strategic planning (top management: long-range goals, new markets, major investments), management control (middle management: budgets, resource allocation, performance monitoring) and operational control (supervisors: making sure specific tasks are done efficiently). Information needs differ by level: operational control needs detailed, frequent, internal and accurate data; strategic planning needs summarised, occasional, often external and forward-looking information. Herbert Simon described decision-making in phases, intelligence (finding the problem), design (developing alternatives) and choice (selecting one), with implementation and review added later. Gorry and Scott Morton (1971) combined Anthony's levels with Simon's distinction between programmed and non-programmed decisions into structured, semi-structured and unstructured decisions.
Transaction processing systems (TPS) record the routine transactions of the business: milk collection, sales invoices, payroll, loan disbursements, stock receipts. They serve operational staff, must be fast and accurate, and supply the data every other system uses. Processing may be in batches (all of a day's collections processed at night) or online and real-time (a bank account updated at once). Management information systems in the narrow sense summarise TPS data into reports for middle managers: scheduled reports (a weekly procurement report), exception reports (only centres whose fat fell below a threshold), demand or ad hoc reports, and drill-down reports that move from a total to its detail. They support structured and semi-structured decisions.
Decision support systems (DSS) are interactive systems that help managers with semi-structured and unstructured decisions by combining data with models. A DSS has three components: a data management subsystem (database), a model management subsystem (model base of financial, statistical and optimisation models) and a user interface or dialogue subsystem; some texts add a knowledge component. What-if analysis, sensitivity analysis, goal seeking and optimisation are typical DSS operations, so a well-built spreadsheet model is a small DSS. An expert system applies rules from a knowledge base through an inference engine, as in pest diagnosis; executive support systems give top managers summarised, often external information with drill-down.
Enterprise systems integrate across functions. Enterprise resource planning (ERP) uses one shared database and integrated modules for finance, procurement, inventory, production, sales and human resources, so that a transaction entered once updates every function; supply-chain (SCM) and customer-relationship (CRM) systems extend integration to suppliers and customers. ERP implementation is treated in Business Analytics and Digital Agribusiness.
Information systems are also a strategic resource. They can lower cost (automated weighing and payment cut the cost per litre collected), differentiate a product (traceability from farm to shelf supports a premium), lock in customers and suppliers (a dealer portal with credit history and easy reordering), and create new business models. Porter's value chain and five-forces models are used to locate where systems add value. Technology competitors can buy gives only temporary advantage.
MIS success depends on management support, user involvement, clear objectives, data quality and training. Systems fail through unclear requirements, design around technology rather than decisions, poor data entry, staff resistance and no one owning the data.
The systems development life cycle (SDLC) is the sequence of stages through which an information system is planned, built and maintained. The usual stages are: preliminary investigation and feasibility study; systems analysis (studying the present system and defining requirements); systems design (logical design of what the system must do, then physical design of how: screens, reports, files, controls, hardware); development or construction (programming or configuring a package); testing; implementation (installation, data conversion, training and changeover); and operation and maintenance.
Feasibility is judged on several dimensions, often remembered as TELOS: technical (can it be built with available technology and skills), economic (do benefits exceed costs, judged by payback, NPV or cost-benefit analysis), legal (does it comply with laws and contracts), operational (will people use it and does it fit how work is done) and schedule (can it be done in time). A system costing 800,000 that saves 250,000 a year pays back in 3.2 years; whether that is acceptable depends on the system's life and the firm's cut-off.
Testing proceeds from unit testing (single program modules), to integration testing (modules together), system testing (the whole system against the requirements), and user acceptance testing (users confirm it meets their needs). Changeover methods are direct or 'plunge' (the old system stops and the new starts on one date: cheapest and riskiest), parallel (both run together until the new one is trusted: safest and most costly), pilot (the new system runs fully at one site, such as one branch or one chilling centre, before roll-out) and phased (the new system is introduced module by module). Maintenance is corrective (fixing errors), adaptive (adjusting to changes such as a new tax rule), perfective (improving performance or features) and preventive (reducing future failure).
The waterfall model runs the SDLC stages in strict sequence, each signed off before the next; it suits stable, well-understood requirements but discovers misunderstandings late. Prototyping builds a quick working model that users try, so requirements are refined by experience. The spiral model repeats planning, risk analysis, building and evaluation in cycles, putting risk management at the centre. Agile methods deliver working software in short iterations, often two to four weeks, with continuous user feedback and acceptance of changing requirements. Many organisations buy and configure packages rather than build: the trade-off is speed and proven functions against fit to local processes.
Requirements are gathered by interviews, questionnaires, observation and workshops. Analysts then model the system. A data-flow diagram (DFD) shows how data move through a system using four symbols: external entities (sources or sinks outside the system, such as farmer or bank), processes that transform data, data stores, and data flows; the context diagram is the top-level DFD showing the whole system as one process with its external entities, and lower levels explode it into detail. Flowcharts show the logic and sequence of a process. Decision tables and decision trees set out complex conditional rules, such as loan eligibility, so that no combination of conditions is missed: with three yes-or-no conditions there are 2^3 = 8 rules to check. Object-oriented analysis uses the Unified Modeling Language (UML).
Agricultural information systems apply all of the above to the needs of farmers, firms and governments. Market information systems collect and publish prices and arrivals in wholesale markets so that farmers and traders can decide where and when to sell; their value depends on timeliness, coverage and the trustworthiness of the reporting. Farm management information systems keep records of fields, inputs, operations, yields and costs and support planning and compliance. Traceability systems link each lot of produce to its farm, inputs and handling, supporting food-safety recall and certification; they rely on unique identifiers, barcodes, QR codes or RFID tags and disciplined data capture at each step. Cooperative and procurement systems automate weighing, quality testing and payment at collection points, reducing disputes and fraud. Advisory services deliver weather, pest and market messages by voice, text and apps.
Governments increasingly build digital public infrastructure for agriculture (farmer and land registries, electronic markets, subsidy payments), which reduces transaction costs but raises issues of data ownership, consent, accuracy of land records, interoperability between departments and exclusion of farmers without devices or documents.
The practical rules hold at every scale: start from the decision, capture and validate data once at the source, store it in a designed database, secure it, and plan for poor connectivity and limited literacy. The final test of an agri-information system is whether a farmer, an officer or a manager takes a better decision sooner than before.
A student uses this subject every week: cleaning survey data and running crosstabs and regressions in a statistical package for a thesis, building budget and cash-flow models in a spreadsheet with absolute references and lookups, summarising market-arrival data with pivot tables, and presenting results. For NABARD and bank officer papers the student must answer computer-awareness questions quickly and exactly: memory types, storage units, number conversions, SQL command groups, protocols and security terms. The habits worth forming now are separating inputs from formulas, declaring missing-value codes, keeping a saved script for every analysis, and never typing a number into a formula.
Five years into a job at a bank, input company, processor or cooperative, the manager owns reports and small systems. Typical tasks: specify an exception report for procurement or overdue loans, write or check a SQL query against the branch or dealer database, build a pricing or credit-appraisal model with scenarios and goal seek, run a user-acceptance test for a new collection-centre application, choose between a SaaS product and a custom build, and handle a phishing or ransomware incident by isolating machines and restoring from offline backups. The decisions are which report actually drives action, what access each user needs, and which changeover method limits risk in the busy season.
With twenty-five years' experience, the executive decides information-systems strategy and investment. Questions at this level: should the cooperative move core systems to a public, private or hybrid cloud, and what are the exit costs; does an ERP or platform investment pass a cost-benefit test once training, data cleaning and process change are counted; how to meet data-protection law across farmer registries and loan files; which security standard and audit to adopt; how to keep an offline or assisted channel for farmers without smartphones. The senior manager also sets data governance: who owns each data set, how quality is measured, and which decisions the board expects the MIS to inform.
A researcher with long experience studies information systems as economic and social institutions. Research questions include the effect of electronic markets and market-information systems on price dispersion and farmers' share of the consumer price, the adoption and impact of digital advisory and payment services, the economics of data sharing and interoperability in digital public infrastructure for agriculture, and the design of DSS that farmers and lenders actually use. Methods include randomised and quasi-experimental evaluation, adoption models and design science. The researcher must also model reproducible practice: version-controlled scripts, documented data, and honest reporting of where systems failed.
This manual teaches concepts and generic functions, not the menus of any particular spreadsheet, database, statistical or office program, because those differ between products and versions; function names and argument conventions should be checked in the software in use. It does not teach programming, network configuration or database administration. Predictive analytics, machine learning, GIS, digital marketing and ERP implementation are covered in Business Analytics and Digital Agribusiness, and the statistics behind package procedures in Business Statistics and Research Methodology. Data-protection, e-signature and cyber-crime law are named only at the level of their existence and general principles; they differ by country and change often, and no section numbers or penalties are given. Market sizes, adoption rates and current platform statistics are deliberately omitted because they change quickly.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.