Data Science 2021

Course Portfolio - updated weekly

Week 1 - Getting Started

Graph taken from Rapidcharts.io. Data fetched from UK gov covid API, so should automatically update in accordance with new daily Covid data.

Week 1 - Embedding

Graph taken from Rapidcharts.io. Data from the ONS measuring long-term UK productivity and hosted in CSV format on Github.

Week 1 - Embedding

Graph from Rapidcharts.io, source code taken from Github and embedded here.

Week 2 - Hosting Data

UK government Covid API with the data filtered to 3 Upper Tier Local Authority areas. This should update daily in line with new data being added.

Week 2 - Hosting Data

UK government covid data downloaded in CSV format from the Covid API and hosted on Github. This process would need to be repeated to show recent data.

Week 3 - Editing Data

This data was downloaded from the Emissions API and pushed to Github. It has been edited on Github to show 3 further (ficticious) observations from the 19th - 21st November. The API takes data captured by the Sentinel-5P satellite and provided by the ESA.

Week 3 - Editing Data

This data was downloaded from the Sports Data API and copied to be 'in-line' within the vega-lite specification JSON file.

Week 4 - API driven charts

This graph uses live data from the Sports Data API. It ranks Premier League teams by the number of unique players to score a goal in the 2021/22 season.

Week 4 - API driven charts

This graph uses data from the Nomics Crypto API which allows a free sign-up. It features 10 of the largest cryptocurrencies during November 2021 (excluding stablecoins). The free API has a requests limit so the graph may not always show up to the minute data. Hovering over the data points gives the time accurate at.

Week 5 - Loops

This graph uses data from FRED, downloaded using a python loop and then uploaded to Github.

Week 5 - Loops

This graph uses data from FRED, downloaded using a python loop and then uploaded to Github.

Week 5 - Loops

This graph uses data from FRED, downloaded using a python loop and then uploaded to Github. The sudden jump in 2020 is due to a change in method for measuring the M1 money stock whereby savings were newly included.

Week 5 - Loops

This graph uses data from FRED, downloaded using a python loop and then uploaded to Github. The sudden jump in 2020 is due to a change in method for measuring the M1 money stock whereby savings were newly included.

Week 7 - Loops and Scrapers

The data for this graph was created by python scraping tidal data from tidetimes.org.uk. The script finds the current day, and then uses a loop to adjust the URL and take tidal data from webpages for each of the next 7 days. This is then cleaned and exporting as JSON to be uploaded to github. This could be developed by auto-pushing the new JSON file everytime the script is run, or having the script auto-run by itself to the graph always shows the next 7 days.

Week 8 - Data Stories
Festival of Economics - Economics of inequality

During this talk, Prof. Jagjit S. Chadha from the NIESR - claimed the UK had the largest fall of outputs (quantity of goods and services produced yearly) during covid than any other advanced economy and one of the slowest recoveries. This graph uses data downloaded from the OECD to show the changes in output for some of the world's major economies, and an OECD average, during the Covid-19 pandemic

Week 9 - Merging data

Ths graph considers the impact on the cost of solar as the capacity is increased. Both axis use a logarithmic scale, with each tick to the right representing a doubling of solar capacity globally. The dataset was created by combining data appropriate data from the IRENA 2021 Solar Cost publicationa and their online Query Tool for global solar capacity.

Week 10 - Advanced analytics

This graph explores if there's any clear correlation between the daily returns of Bitcoin and Gold. I initially used data from MetalsAPI but due to some missing data points in the daily pricing, I switched to using PAXG, a crypto stablecoin for gold, with data obtained in csv from coinGecko. The BTC data was obtained from the Nomics API and combined with the PAXG pricing data using Python. I then calculated their simple daily returns and exported in JSON format to be used in Vega-lite. The result shows no clean correlation between assets, however is not conclusive as it only considers returns at the daily frequency.

Week 11 - Interactivity

Using data from Sports Data API, this graph filters for the current Premier League Season and shows the gold scored for the home team in every match this season. Click any result to isolate that team's home results, double click to deselect.

Week 11 - Interactivity

Area chart uses data from IRENA using their VBA excel data query tool for Renewables Capacity and Generation. This was then exported as a CSV file and uploaded to Github. Use the sliders to explore the full data period is 2000-2020.