AI Micro-Internship · ~180 min · low-code

Make a real forecast — then face the actuals

Anyone can analyse the past; forecasting makes you accountable to the future. Build a simple forecast from historical data — tomorrow’s temperature, AQI, vegetable prices, petrol rate — write your predictions down BEFORE, then verify against reality for 3 days and score your error.

Steps

01Pick something with public daily values and get 2+ weeks of history (weather sites, AQI portals, fuel price trackers).
02Build a simple forecast method: moving average, same-day-last-week, trend line — AI can help, but you must be able to explain it.
03Write down predictions for the next 3 days WITH dates, before the days happen.
04Each day, record the actual value. Compute your error (predicted vs. actual) and compare against a naive baseline ("same as yesterday").

Tools: A public daily data source, Google Sheets, Three days of patience

Your submission

Submit your method, the dated predictions, the actuals as CSV, and the error analysis vs. the naive baseline. Honest misses beat suspicious accuracy.

What you forecast, the history you used, and your method explained in your own words *Your 3 dated predictions, exactly as written before the days happened *CSV: date, predicted, actual, error — plus the naive baseline’s error *The analysis: your error vs. naive baseline, where you missed and your theory why *