My Experience With Flatiron School’s Immersive Data Science Bootcamp — Part 4

Week 3 of Data Science marked the beginning of the fun part of my data science journey. I detailed how in Week 2 things became easier for me once I realized I could handle SQL. I started to get into the swing of things and genuinely enjoyed the work we did.

This week was all about data visualizations, and learning how to communicate our results. I would say this was the easiest week for me. My past work had left me comfortable explaining complicated topics to lay people. I had a method, called the reverse pyramid, where I would start broadly with a topic, to introduce background or setting, then piece by piece, focus in on the topic I wanted to address, or the point I wanted to reinforce. This way, people had both the proper perspective, and an understanding in how I got there. I’d finish with conclusions, and then some useful but optional info. This has worked spectacularly in the past with clients i’ve had, on topics ranging from backdoor Roth IRAs to Charitable Remainder Trusts.

So about data visualizations. We started with learning how to use matplotlib, and its add-on, Seaborn. We learned how to make things like pie graphs, bar charts in all its myriad forms, as well as line graphs. We were then given the freedom to go through the documentation and make as many bizarre graphs as we could using all the various extensions each type of graph matplotlib could afford us.

Seaborn, I liked best, it made nicer graphs, was easy to use, and had way more customization options as compared to matplotlib. Its largely built on matplotlib, but its presentation style is easier on the eye, cleaner, and has a more professional looking color palette. For example:

Matplot lib on left, Seaborn on Right. Credit: Codeyarns.com

Our project for this week, was to incorporate the previous 3 weeks of lessons and put together a presentation where we were to pitch Microsoft on what direction they should take their streaming service. Should it be a movie or a tv show? What genre? What actors should they use for it?

I was partnered with 2 classmates, and we came up with the idea for “MicroFlix,” Microsoft’s new streaming service. The idea was to target Microsoft’s already locked in customer base, X-Box Live users. It would be so easy to build a streaming service into X-Box live, because Netflix and other streaming services already lived there. You could just drag eyes from those other streaming services with content already lined up for that demographic.

Using Microsoft’s own XBox demographic data, we discovered that the majority of XBox owners were 52% male, and overwhelmingly in the 18–54 age range, with an average age of 33 years. Then we went looking for data that would tell us “What do these people like to watch?” For that, unfortunately, we had to pull from Statista, because apparently demographic data is not free. Those marketing firms hold on to their research data pretty tightly.

Statista told us what we pretty much knew already, that demographic loves their Action, Adventure, and Comedy movies:

So what genre were we going to pick? Action? Adventure? Comedy? Well to answer that, we had to start doing real work.

We visited the-numbers.com, a fantastic resource for movie data and scraped their site for genre data. We were attempting to find trends in gross numbers that would help us pinpoint a genre to start with. Here’s what we came up with using Matplotlib:

Yes, this is matplotlib. See how much it sucks?

Action movies have seen a steady increase in market share due in large part to the rise in comic book movies. Studios like Marvel were churning out epic sagas (good ones) that just dominated the box office. Adventure movies also owned a large part of the market share, but had seen a dip recently against action. Comedies, though largely favored by many in our targeted demographic just didn’t bring in the same amount of money, no matter if it were just straight comedy, black comedy, or rom com.

So we knew comedy was out. It could be great, but wasn’t going to make the splash that would do well as Microsoft’s debut into the streaming space. Did we want to hop on to the Action genre, and coast along on Marvel/Disney’s coattails? Sure it might be a sure fire win if we got the right actor and story for it. Maybe Michael Bay it up a bit for that extra pop? The thing is, my team and I were worried about genre fatigue. There were so many big action films out there now, we knew there was a guaranteed audience out there for the next Marvel or Star Wars flick, but that wasn’t an audience you could guarantee would be interested when an also ran stepped on to the field. Case in point, Warner Brother’s attempts to get the DC universe into that money making field. The Dark Knight series were amazing, and Wonder Woman is a popular contender for greatest comic book movie ever made, but everything else just felt too….forced. As much money and cinematic talent goes into that universe, everything felt like they were screaming for attention with maybe only half of what Marvel could offer viewers. People were getting tired of it. We felt like if you were going to step up to Marvel, you either had to step it up big in terms of money and talent, and it would be a lot of work for Microsoft.

So we decided to side step it, thinking that Action’s market dominance was soon approaching its peak, and took a look instead at Adventure, which was on its way back to recapturing market share. What kind of Adventure film? Well Microsoft had also captured data on the gaming preferences amongst their owners, i’ll share that below:

We can see some definite spikes toward solo (vs. teamwork), and towards nostalgia. This makes sense for an average age of 33 demographic. So we’d want something with a focus on a single hero, rather than an ensemble cast, and we’d want something that is a revitalization or reboot of an old franchise. Indiana Jones comes to mind, or even a film with an homage to nostalgia, like Ready Player One.

The rest of the project followed down these paths, where we attempted to isolate for a specific set of directors or actors to appear in our new film. We used API calls from themoviedb.org and omdb.com to determine popularity rankings for major actors by genre. Elizabeth Shue, Sean Bean, and Jackie Chan are up there, along with a few foreign actors we weren’t too familiar with. For a directors, by far the most popular were Joss Wheedon, Darren Aranofsky, and Kenneth Branagh. I’m not familiar with all their works, but I strongly feel like a Sean Bean/Joss Wheedon pairing would really generate some buzz.

All in all, this was a fun week. I got to work on projects and presentations, which I thoroughly enjoy. The subject matter was interesting, although that necessarily wasn’t that important. Most importantly, I put a lot of the things I struggled with in the past two weeks to use, and understand how it all comes together to influence a business decision. Next week: The Math Starts!

Data Scientist, Financial Planner. Trying to educate and make information accessible to EVERYONE. Let’s Connect! shorturl.at/aBGY5