Introduction
When looking at timeseries data, we may encounter high variability from one week to the next, creating highly jagged and difficult to read charts.
There are a number of techniques to smooth the lines to better depict trend.
Some may use rolling averages, others, like YouGov, use LOESS regression.
https://yougov.co.uk/topics/society/trackers/how-often-brits-exercise
In this weeks challenge we utilise Deneb to create a Loess regression. Collect your data from either YouGov, Financial Times or another datasource of your choice and see if you can recreate and enhance the charts.
EXTRA POINTS for line brush markers.
Don’t forget to cite your source.
Happy Vizin’
Kez
Requirements
- Obtain the data
- Obtain Deneb from Appsource
- pivot the category data with Power Query or use folding transform in vega-lite
- create a scatter plot coloured by category
- create a second layer
- using a transform function create a LOESS regression on date
- use the LOESS transform output to create a lines coloured by category
- format your chart and add desired amount of interactivity
Dataset
This week’s dataset is downloadable from The Financial Times OR YouGov
Share
After you finish your workout, share on Bluesky or LinkedIn using the hashtags #WOW2025 and #PowerBI, and tag @MMarie, @shan_gsd, @KerryKolosko. Also make sure to fill out the Submission Tracker so that we can count you as a participant this week in order to track our participation throughout the year.
Solution
Coming soon.