IoT Sensor Data Analysis: Resampling, Thresholds, Drift & Uptime

What you'll build

What you will be able to do

Each lab ends with a concrete result that you build yourself and the checker confirms, cell by cell.

• Turn the timestamps of a cold-room sensor into intervals in minutes, flag the gaps, count the missing readings and find the longest outage.
• Put 15-minute readings into hourly buckets and compute readings per hour, hourly mean and hourly maximum, the way dashboards downsample sensor data.
• Flag every reading outside 2–8 °C, add up the minutes out of range, measure the longest excursion with a running counter and count separate excursions.
• Count received and valid readings for four sensors, compute data completeness and compare it with the uptime from the gateway log.
• Compare twelve weekly spot checks with a calibrated reference probe, measure the drift per week with SLOPE and predict when the sensor leaves its tolerance.
• Smooth 5-minute readings with a 30-minute moving average, compare how many raw values and how many averages cross the alarm limit and chart both.
• Build a step table for a simple threshold alarm and a hysteresis alarm and count how many notifications each one would send.
• Turn two weeks of daily sensor summaries into a compliance report: a status per day, the days OK, the time in range, the mean kinetic temperature and a chart of the daily minimum and maximum.