Families
Predict
Estimating a value you do not have yet, or flagging that an observed value is out of the ordinary. This is the family that demands the most data and is easiest to get wrong, because a plausible prediction is hard to tell from a correct one. A simple method you can evaluate beats a clever one you cannot check.
What is going to happen, and what looks wrong?
2 entries
- Forecast next week's sales RecommendedN1
- Spot an anomaly in a metric RecommendedN0
Roadmap
This need is missing from the catalogue. Propose it, that is how it grows.
- Announce a waiting time Propose this entry
- Anticipate a stock shortage Propose this entry
- Check that a model is not drifting Propose this entry
- Estimate a delivery's carbon footprint Propose this entry
- Estimate a second-hand price Propose this entry
- Estimate a text's reading time Propose this entry
- Estimate attendance at an event Propose this entry
- Estimate how long a job will take Propose this entry
- Estimate the no-show rate for a slot Propose this entry
- Fill a gap in a series of readings Propose this entry
- Find the seasonality in a series Propose this entry
- Forecast site load Propose this entry
- Forecast staffing needed for a shift Propose this entry
- Forecast the coming weeks' cash Propose this entry
- Give a delivery date Propose this entry
- Plan maintenance before the breakdown Propose this entry
- Put an uncertainty range on a forecast Propose this entry
- Spot a lasting change of trend Propose this entry