Ever wondered where in London most start-ups are born?
Well, you’re in luck! We will plot an animated graph of exactly this.
(Or skip to the bottom to see it now)
A jupyter notebook containing code for this analysis can be found here.
To obtain information about company headquarters, we have scraped the REST API of Companies House, the UK’s registrar of companies. Their API (link here) is outstanding, and commendable for supporting public transparency and open-data standards.
We scraped 10K companies with listed offices in London and that had been founded in 2011–2020 using said API and a python script using the python request package. We recorded the company ID, date of creation, date of cessation of the company (if it had been closed down), and the company registered office postcode.
Next, we simplified the postcodes to the leading area codes, and created a numpy matrix of active companies per postcode in the 109 months between 2011 and 2020 we were interested in.
We decided to use the plotly python library for plotting choropleth map views of London across the timeline discussed. To plot on a map using postcodes, we needed a GEOjson file that specifies the latitudes and longitudes of the postcode areas, which we obtained from the GitHub user sjwhitworth (from here).
Finally, we simply looped over the months from 2011 to 2020, and saved PNG outputs, which we converted to GIF-FI format using the imageio library.
The graph shows the square root scale (to not skew the graph too much by the most active postcodes) of net active companies per postcode region, per month. The outskirts are not recorded due to their postcodes residing outside inner London.
Thanks for reading, you may enjoy our other articles as well:
Can you Time Economic Downturns?
The market can stay irrational longer than you can stay solvent. ~ Keynes