In Knarr CTO Speros Kokenes’s latest Data Diary, he explores data from the Economic Impact Tracker released by the research and policy analysis organization Opportunity Insights. Their tracker includes various economic indicators related to COVID and which states and areas have been affected most severely. Some metrics include commercial spending and small business revenue.
This tool is an interesting way to explore this data in predefined ways, but since the organization has compiled all their data in the open-source repository GitHub, we can explore the data freely by pulling it from the repository and pulling it into Knarr. The repository includes geographic information related to US localities, spending averages over time, job posting data, business revenue, and more.
To begin exploring this collection of economic data in Knarr, Speros first creates and names a new project in Knarr and begins a Note to document the source of the data. He then downloads the data to explore from GitHub and converts it to CSV files as needed to then upload it to Knarr. Speros’s exploration begins by examining the Burning Glass job posting data.
Once the job posting data is loaded you can see that this data only includes codes for the US states it corresponds to. This is where Knarr’s ability to link datasets comes in handy as Speros pulls a different geographic dataset to link the FIPS codes to state names for clearer analysis using state names and the jobs data.
First Speros creates a line chart showing job postings over the past few months including an average of state indices of job postings. Then he creates a national job post trend table to compare with the average postings by individual states. This is useful to see how job trends vary from state to state– indicating that most are following the trend of postings dropping significantly in April then having an upturn in June. By then ordering the state grid according to overall value ascending– the states with the worst changes move to the top of the list and the least affected states are at the bottom. This sorting reveals that some states like New Hampshire have actually remained strong throughout this period.
This data is useful for comparing states and determining which states are doing better than others economically. To continue the state by state economic comparison, Speros pulls in spending data to create a comparison of overall spending indices by states then isolating states hardest hit by jobs and comparing those. We then see how jobs data doesn’t exactly correlate with spending information. In this period spending has actually gone down the most in Rhode Island, DC, and California. Conversely, spending trends have leveled off in other states like Tennessee, Iowa, and Arkansas.
Another interesting way to examine this data that’s available is to compare different income sectors across the country. One assumption is that spending likely varies by household income brackets. An interesting insight that this sorting shows is that the trend was sharper according to ascending income brackets with a sharper trend among higher income brackets. This can probably be explained by higher-income households having a greater level of disposable income and thereby being more able to suspend unnecessary spending with a smaller income percentage being used for essential items. Using the grid feature, Speros then breaks this trend out to view individual states and note differences in how these economic trends have affected them.
Lastly, Speros creates a scatter plot comparing state job posting changes and average spending changes. For additional detail in navigating he creates a side-by-side bar chart showing states and average job postings sorted by number. By highlighting the state of New Hampshire– the only state to show rising job postings throughout this period– it’s clear that this state is also an outlier when it comes to spending, as shown on the adjacent scatter plot.
Using Knarr’s Timeline feature, Speros is then able to go back in time in the analysis to pull up the previous chart he made of spending by income brackets. By then modifying that chart to look at overall spending (not segmented by income level) and then adding a metric showing job changes, you can easily see the correlation between these indices in individual states. This analysis certainly raises questions about different public policies and how the current pandemic is affecting businesses and livelihoods differently across the country.
While we are all feeling the impacts of the COVID-19 public health crisis, it is a different picture in different places. Hopefully, through analysis and monitoring, we can learn lessons from this difficult time and apply these lessons as we move forward in the future of our country. As we continue on in a rapidly changing world, we hope Knarr can be useful to those of you exploring the data that accompanies our trajectory. To explore more for yourself, visit us at knarr.io to start your journey today.
Header Image from https://www.tracktherecovery.org/, July 2020