Welcome!
Data is an extremely powerful tool. Most of the examples you'll see here comes from personal projects. Non-Disclosure Agreements are often tied to the work we do as Data Analysts, Scientists & Strategists. Don't hesitate to reach out!
[email protected]
Christian's LinkedIn Account
Thanks for stopping by! My name is Christian Acosta. That's my beautiful wife Caitlin and our four-legged menace, Penny. We got married 9/12/15.
I began studying Data Science in January 2016 at the University of Wisconsin. I'm almost done - can't believe it!
I owe my inspiration to Caitlin who encouraged me to pursue my dreams of being an entrepreneur and data scientist. Without her support, help and guidance I would not be the man I am today.
I began studying Data Science in January 2016 at the University of Wisconsin. I'm almost done - can't believe it!
I owe my inspiration to Caitlin who encouraged me to pursue my dreams of being an entrepreneur and data scientist. Without her support, help and guidance I would not be the man I am today.
A Few Things I've Worked On Recently
|
Network Visualization of a (primarily) Women's Apparel Brand's Web Domain
The larger the circle, the higher the Google PageRank. Red is dedicated to men's products. Purple is dedicated to women's products. Green is the home page. The pages with the highest page rank are all men's products! Each line (radiating outwards) is a link to another page. We immediately notice that the domain is very dense. There are a handful of pages with a ton of links, and a handful of pages (on the outside) with very few links. An intelligently-linked website ranks higher on Search Engines. Google's page ranking algorithm takes three things into account: 1. How many sites are linking into a given page. 2. How many links are going out of said page. 3. How popular the page is - does it have other popular pages linking to it? |
Predicting if a Basketball Shot Will Go In
Wearable technology is all the rage. What if we could use the player's momentum, court location and opponent's location to increase shot selection quality in NBA players?
The picture shows the average field goal percentage (100,000+ shots) in the NBA for the 2014-2015 season. Unfortunately my results weren't all that compelling, but let me bore you anyway!
I used K-Nearest Neighbors, Logistic Regression and a Neural Network. The Neural Network predicted 61.4% of the shots correctly ... BUT WAIT!
The model correctly predicted that a shot was missed 87% of the time! Nice. However, it only got 31% of the ones that actually went in correct.
Should we make that wearable? It depends if you believe in negative reinforcement.
Wearable technology is all the rage. What if we could use the player's momentum, court location and opponent's location to increase shot selection quality in NBA players?
The picture shows the average field goal percentage (100,000+ shots) in the NBA for the 2014-2015 season. Unfortunately my results weren't all that compelling, but let me bore you anyway!
I used K-Nearest Neighbors, Logistic Regression and a Neural Network. The Neural Network predicted 61.4% of the shots correctly ... BUT WAIT!
The model correctly predicted that a shot was missed 87% of the time! Nice. However, it only got 31% of the ones that actually went in correct.
Should we make that wearable? It depends if you believe in negative reinforcement.
If Engagement's Up, Then Where's the Revenue?!?
Imagine you pay thousands of dollars a month to have your brand's Facebook account managed by social media professionals (check out that out-of-place black arrow). They claim your Facebook Reach (total views) & Engagement (likes, comments, shares) on your brand's Facebook page has hit an all time high... but you tell them your revenue is completely flat. Confusing? Well - that's what one of my clients was going through.
So what would you say... Ya do here social media company?
Digital Marketing is all about brand-to-consumer engagement. Having people see your posts is great. But are people engaging with your content? I created this plot using R to pull down public data from Facebook. Immediately we found the source of their problems - people were liking the pretty pictures but they weren't buying anything.
The solution is simple: start speaking with your customers instead of speaking at them!
Imagine you pay thousands of dollars a month to have your brand's Facebook account managed by social media professionals (check out that out-of-place black arrow). They claim your Facebook Reach (total views) & Engagement (likes, comments, shares) on your brand's Facebook page has hit an all time high... but you tell them your revenue is completely flat. Confusing? Well - that's what one of my clients was going through.
So what would you say... Ya do here social media company?
Digital Marketing is all about brand-to-consumer engagement. Having people see your posts is great. But are people engaging with your content? I created this plot using R to pull down public data from Facebook. Immediately we found the source of their problems - people were liking the pretty pictures but they weren't buying anything.
The solution is simple: start speaking with your customers instead of speaking at them!
|
Do MLB teams with a higher number of Twitter followers pack their stadiums fuller?
The larger & greener the circle, the higher the winning percentage. Redder & smaller? No bueno. See that yellowish dot in the top right? That's the S.F. Giants. They almost always fill their stadium & have the highest ratio of twitter followers when compared to the size of their TV market at 0.35'ish. Say what? For every 10 people in the San Francisco TV Market, there are almost 3.5 people following the Giants on Twitter! Actually, having a high Twitter-to-TV-market ratio is a highly-significant relationship with getting a packed stadium! This suggests that a strong Twitter presence can lead to fuller stadiums. BUT, you guessed it, those teams happened to win more often. So, is Twitter the key to packing your MLB stadium? Probably not. But, do winning teams attract Twitter followers? Hmmm |