Skip to content

joeledwardson/business-activity-visualiser

Repository files navigation

Business Activity Visualiser (backend)

Backend of business activity visualiser, see demo deployed here

Back-end API is served together with the frontend

Data cleansing

The data folder contains the Python for for data processing and cleansing from the ONS business activity statistical report from 2021.

The excel file holding the data downloaded is data/ukbusinessworkbook2021.xlsx.

The Jupyter notebooks do the following:

  • data/datageography.ipynb retrieves and processes ONS geographical data for each region
  • data/datasheets.ipynb retrieves and processes data from each sheet in the excel workbook identified above.

The result of both notebooks is producing data/businessdatabase2.db, an SQLite database holding the parsed data from the workbook, and geoJSON data for each geographical entity.

Because the spreadsheet geographical codes are a mix of countries, counties and others, a level system is created so that no selection of geographical areas is duplicated

Level Name Notes
1 UK
2 Great Britain (GB) and Northern Ireland (NI)
3 (England and Wales), Scotland, Northern Ireland England and Wales as a single entity
4 County England and Wales separate countries
5 Region
6 County Mix of counties, London boroughs, Unitary Authorities
7 District

Server

The root project entry point is main.js, a node server providing an API to the frontend to retrieve numeric and geoJSON data from data/businessdatabase2.db (and frontend React application.)

requests

Requests are accepted in post format to <host URL>:8000/api/geocodes with:

  • level as an integer, between 1 and 7

Returned is a JSON object with either an error:

  • error contains error message

or and object:

  • message: "success"
  • data: a geoJSON FeatureCollection of each geographical entity at that level, for each feature the properties attribute has keys of industry group from table 1 and the value is the summation for that geographical area

Querying

To view the frontend complete deployed application, navigate to: https://business-activity-visualiser.herokuapp.com/

To access the underlying API, can use the /api/geocodes URL:

Future Work

At the moment the server is hard coded to use table 1, whereas the database contains data from all 26 tables presented in the spreadsheet.

Some tables do not have geographical information so rendering on a map would be futile, but for the other tables that do a frontend selector could be created.

However, the geographical information is sometimes in the sheet rows and sometimes in column headers, so this would need to be indicated to the frontend.