D4 Capstone Scope, Field Validation, and Farm Data Privacy
A D4 Electronics Engineering capstone team of four students has a 12-week project with a partner SME. The SME wants a low-cost soil-moisture sensor for 50 smallholder farms; the budget is IDR 25 million; the team has 10 lab hours per week; one member has a 6-week internship conflict. The partner demands a working prototype and field data from at least 20 farms. A simple calibrated resistive sensor can be built in 4 weeks and tested on 8 farms within budget; an IoT system with a cloud dashboard would impress the partner but needs 10 weeks and leaves little field testing. The partner also wants raw farm location data, which raises privacy concerns. The team must present a scope decision at week 2.
1. Decide which technical scope the team should propose and justify the trade-off between technical ambition and field validation.
Read the worked answer
Propose the simple calibrated resistive sensor with an expanded but still feasible field plan of 10–12 farms, not the IoT system. The partner's core need is a working low-cost sensor, not a dashboard. The IoT system needs 10 weeks, leaving only 2 weeks for field testing, and one member is absent for 6 weeks. With 4 weeks for build and 6 weeks for testing, the team can deliver a prototype, a calibration protocol, and a farmer usability report. Negotiate the 20-farm target down by explaining that 10–12 farms can detect gross calibration error; 20 farms adds little if the sensor is not stable. Use the saved budget for spare sensors, transport, and a second calibration round. This scope protects validity and deliverability, and it gives the partner a tested product rather than a fragile demonstration.
2. What risk and privacy measures should accompany the proposal?
Read the worked answer
Risk measures: assign the internship-conflicted member a remote data-analysis role with weekly deliverables; freeze the hardware design at week 4; run a fail-fast calibration test at week 2; keep a backup sensor design; and schedule field visits around the absent member's availability. Privacy measures: collect only farm-level anonymised IDs, not GPS coordinates; store any raw location data on a password-protected device; share only aggregated results; and state in writing that location data will be minimised to comply with data-protection norms. The team should also report sensor accuracy with confidence intervals and explain that the 10–12-farm sample limits generalisation. These measures address the partner's demand for field data without creating unnecessary privacy risk or overclaiming statistical representativeness.