I would like to submit an entry for Dec 2023 Presentation (Development time: ~3h).
Link: Google Colab Jupyter Notebook
I wanted to explore alternative visualisation techniques for plotting multivariate time-series and to move beyond always having my time dimension on Y-axis. I also wanted to experiment with darker themes.
Dashboard connects to FRED API using pandas-datareader and to Bank of England database through requests to fetch data. The dashboard plots monthly money supply (total money in circulation) in US and UK currencies and overlays it over daily FX Cable rate . Dashboard also calculates correlations.
I kinda thought about FX exchange rate as a ratio between money supplies (total money in circulation):
- If Money Supply of GBP increases, while Money Supply of USD stays the same - Dollar appreciates
- If Money Supply of GBP decrease, while Money Supply of USD stays the same - Dollar depreciates
(all else being equal)
- RangeSlider for Time (1y, 5y, 10y tabs) is only available for xaxis, so if you are switching orientation of plot, you would not be able to use RangeSlider.
- I wanted to add correlation table and settled down on dbc.Table with html Threads, recycling code from one of my previous projects. I have individual output for every correlation value so it’s not an optimised way to write a callback and I would suggest to return a list.