Categories
Dashboards Data Tableau

Fails to Deliver and GameStop – A Look Inside

Click here to jump straight to the dashboard

You probably haven’t heard anything about GameStop in the past few years. Nope, nothing. Just a defunct retail chain where you used to be able to trade in your used video games for $3 a pop.

Sarcasm aside, the world was taken by storm a few weeks ago when GameStop stock went on a rally. With an exceptionally meteoric rise paired with high short interest, panic ensued for short sellers and smaller brokerages. This led to a rise in failures to deliver. In the most basic definition, this is one of the parties in a transaction did not deliver cash or the asset before the settlement date of the transaction.

Check out Investopedia’s definition below, and definitely read the page if you want to learn more about failures to deliver.

Whenever a trade is made, both parties in the transaction are contractually obligated to transfer either cash or assets before the settlement date. Subsequently, if the transaction is not settled, one side of the transaction has failed to deliver. Failure to deliver can also occur if there is a technical problem in the settlement process carried out by the respective clearing house.

https://www.investopedia.com/terms/f/failuretodeliver.asp

The SEC publishes data on failures to deliver twice every month. At the time of this posting, they had data released up to January 15th, 2021 (update: on Feb 16th I updated this with the SEC data up to January 29th, 2021). This unfortunately doesn’t capture the most extreme part of the GameStop stock price explosion, but captures other volatile periods in the stock’s history. I’ll update the visualization when the data becomes available.

In this dashboard, I wanted to see the correlation between price increases and fails to deliver. Also I was interested in seeing what the comparison was between the average of all stocks reported fails to deliver versus GameStop. Check out the visualization below to see what I came up with. It’s interactive and best viewed on desktop (but has a mobile configuration as well).

The Dashboard

p.s. if you know what stonks are then you know it’s not a misspelling. If not, don’t worry about it and enjoy the ride 🚀🚀🚀

One last note is that this project was a great way for me to practice my Python skills. SEC.gov provided the files in two week chunks, each as a zip file. I made a quick Python script to download all of the data for the last year, unzip the folders, append all of the text files to each other, and output a nice csv to use in Tableau. Python is a semi-frequent skill for me so it’s always nice to have a quick touch-up project like this.

If you’re looking to learn Python, a project like this is a simple way to learn a real world use. Don’t be afraid to jump into projects like this. Break it down into tiny steps and complete one at a time.

Thanks for reading. Any questions or comments? Feel free to reach out to me via email on my contact page.

Categories
Dashboards Everything Else Tableau

2020 Presidential Election: Florida early and mail-in voting participation

Election day is closing in! Let’s take a look at one of the swing states current voting stats: Florida. Florida releases early voting and mail-in voting statistics on their Division of Elections site. I used that data to build a visualization of current trends.

I’d recommend using the visualization below by opening the 2016 Presidential election results (Trump vs. Clinton) and comparing the county results versus the voter participation so far for each party. Voters aren’t guaranteed to vote strictly along party lines, and small deviations from parties can have a big impact in small margin states like Florida. That being said, it can be telling to compare the current votes for each county by party to the results in 2016.

This next part is using numbers from October 30th. For example, Miami-Dade in 2016 had 623,006 votes for Clinton and 333,666 votes for Trump. Currently Democrats have 352k votes and Republicans have 269k votes in 2020. 225k unaffiliated voters have cast a ballot as well. This could mean several things (if we irresponsibly assume people 100% vote along party lines):

  • Republicans have already voted around 81% of their vote total from 2016.
  • Democrats have already voted around 57% of their vote total from 2016. Democrats have about 200k more registered voters in Miami-Dade, but if their participation doesn’t increase significantly, this could indicate an overall negative change in demographic voting for the party across the state.
  • If Republicans keep voting at this rate and the county gets around 75% voter participation, Trump will significantly outpace his 2016 total for Miami-Dade. In 2016 Clinton won Miami-Dade county by a margin of 29 points, but decreased Democratic turnout could narrow that margin significantly.
  • It still all comes down to the voters with no party affiliation. These voters make up 26% of registered voters in Florida, and nearly 33% of voters in Miami-Dade county.

With that in mind, take a look at my viz below! Make sure to hover and click to see additional information.


Interested in getting more posts like this? Subscribe below (I never share my email list with any 3rd parties).

Processing…
Success! You're on the list.
Categories
Dashboards Tableau

Where Did PPP Loans Go? Taking a Deeper Dive.

Recently, I published a dashboard that took a broad look at PPP loan data. Over the past few weeks, I looked deeper into that data set in order to pull more interesting findings from it.

Before you take a look, here are the most significant highlights in my opinion:

  • Spelling mistakes were prominent. Out of some of the 9 largest cities in the USA, there were more than 100 spelling variations.
  • About 4000 investment-related firms took out loans totaling between 1 and 3 billion dollars. This is interesting because investment fees and activities were at record highs during lockdowns.
  • The top 10 industries by loan count were heavily in-person industries, indicating that loans were distributed largely to businesses who needed them most.

While the dashboard has been developed for mobile as well, the viewing experience will probably be better on desktop.

Like this post? Enter your email below to subscribe for more visualizations and Tableau related posts.

Categories
Dashboards Tableau

Looking at PPP Loan Data – A Visualization

$667 billion dollars were set aside in 2020 for businesses across the US. This program was called the Paycheck Protection Program (PPP for short). It was meant to be a stimulus package to hold over businesses who were impacted by COVID-19 lockdowns. Without revenue streams, many businesses had to lay-off workers en masse, providing a potentially catastrophic destabilization of American society. According to the Bureau of Labor Statistics, there were close to 18 million unemployed people in June of 2018. The PPP data indicates that over 31 million jobs were retained due to the program.

Without getting into the inaccuracies of the standard unemployment numbers provided by the Bureau, we can still see that laying off potentially 10% of a country’s population wouldn’t be healthy for society. In the past, the US government has provided stimulus and bailouts primarily via public spending increases and buying toxic assets from big corporations. But this stimulus was meant to directly support businesses of all sizes who were directly impacted by the lockdown. Did it server its purpose? You can be the judge of that! I’ve created a data visualization meant to help show where the money went across the nation. This visualization includes the ability to search and filter on columns of the data. In the near future, I’ll be releasing a dashboard that digs into the… more suspicious loans taken out.

Here’s the dashboard, it’s interactive and has some additional info when you hover over data. It is configured for desktop and phones, but is better to view on desktop.:

This data was recently released regarding PPP loans. It was split into different data sets. For loans greater than $150k, data for the whole nation is provided in one data set. For loans less than $150k, data is split by state or territory into its own data set. I’ll be providing a combined data set of all loans here in a little bit, but for now this visualization is for loans greater than $150k.

Some interesting takeaways from this visualization:

  • Allegedly nearly 10% of the US population’s jobs were retained due to the program
  • There were TONS of misspellings and bad data entry
  • The majority of loans were in the $150k-$300k range

P.S. Check out the next visualization here, where I take a deeper look at the PPP data.

Like this post? Subscribe to be notified of future dashboards and Tableau tips.