ICAR AICE-JRF/SRF Computing, Data & Digital Tools for Dairy: practice questions

1 · Spreadsheets for plant and laboratory calculations · hard

Audits of operational spreadsheets have repeatedly found that:

1. Errors occur only in sheets written by beginners and never in expert work
2. Spreadsheet errors are always caught by the software's built-in checks
3. Spreadsheets used for payments are legally exempt from data-integrity rules
4. A large share contain errors, many of them material

2 · Databases and data management · hard

Which data-quality dimension is violated when the same product has different codes in the production and sales systems?

1. Timeliness
2. Consistency
3. Accuracy of a single measurement
4. Completeness of a single record

3 · Introductory programming and scripting · hard

Why may 0.1 + 0.2 not compare exactly equal to 0.3 in many programming languages?

1. The computer rounds every number to the nearest integer before adding
2. Addition is carried out in base 16 internally
3. Decimal numbers are stored as text strings
4. Binary floating point cannot store most decimals exactly

4 · Introductory programming and scripting · hard

Which practice lets a team see who changed a calculation script, when and why, and restore an earlier version?

1. Version control
2. Data compression
3. Screen sharing
4. Spell checking

5 · Automatic milk collection units and farmer payment systems · hard

An analyser reads fat 0.1 percentage point too high. A centre collects 2,000 kg of milk a day, paid at 5.00 currency units per kg fat. How much is overpaid per day?

1. 10 currency units
2. 100 currency units
3. 1 currency unit
4. 1,000 currency units

6 · Enterprise resource planning in dairy companies · hard

Most failed ERP projects in food companies fail mainly because of:

1. Defects in the database engine that corrupt all records at go-live
2. The cost of computer hardware exceeding the software licence
3. Organisational issues: processes, master data and training
4. Regulations that forbid integrated software in food plants

7 · Laboratory information management systems · hard

Integration of the LIMS with procurement software allows:

1. Analysers to calibrate themselves without any reference samples
2. Farmers to change their own test results
3. Quality results to feed payment automatically
4. The laboratory to skip sample registration

8 · Traceability and recall software · hard

A mock recall traces a butter lot and accounts for only 95% of the quantity produced. The correct response is to:

1. Investigate the gap as a failure
2. Accept it, since 95% is close enough for a perishable product
3. Repeat the test until it closes
4. Change the produced quantity in the records

9 · Data analysis and visualisation · hard

Calibrating a near-infrared analyser to predict fat in milk powder from hundreds of correlated wavelengths usually uses:

1. Partial least squares regression
2. A chi-square test
3. Simple linear regression on one wavelength
4. A Pareto analysis

10 · Sensors, IoT and cloud data in dairy · hard

Edge computing in a remote chilling centre is valuable mainly because it:

1. It removes the need for any sensors at the centre
2. It stores all data permanently on the cooler
3. It increases the cooling capacity
4. It works locally when the network fails