Read on Twitter
Alright, y'all got me. R1D1 #30DaysofThreads
[Thread] How I got into Data Science + How you can Get Started
tbh it was a little luck. In 2014 I was doing social media for a marketing agency. I had 20+ clients like Hilton hotels etc. One day our PPC guy was out sick
[Thread] How I got into Data Science + How you can Get Started
tbh it was a little luck. In 2014 I was doing social media for a marketing agency. I had 20+ clients like Hilton hotels etc. One day our PPC guy was out sick
We had a HUGE client who wanted to drop 35k on a PPC (pay per click) campaign and I was second best suited to help. In a day I had to look at our past marketing click through rates, understand our customer lifetime value, and know which target demograpics to go for.
I had been writing social media content and had no clue what I was doing. I had to recall every nugget our PPC dude dropped about his decisions and add that to my knowledge of the brand to help our client allocate their budget. A big holiday was coming up so it was time sensitive
I helped them target 2 demographics, the out of school teens and stay at home moms. They spent their entire 35k banking on these ads & made 3x over that holiday weekend than the year before. What does that have to do with Data Science? That was my first time dealing w/ analytics
While at this agency I took a bigger interest in our paid campaigns. I'd sit down with our PPC dude and have him explain his choices & what all the metrics were.
That exp helped me land a Social Media Analyst job at an app company. Here my eyes really opened to product data.
That exp helped me land a Social Media Analyst job at an app company. Here my eyes really opened to product data.
I got to deal with 150k real users and look at all different kinds of events. I combined our product trends with our social media trends to have a baseline to A/B test our in-app and push notifications. This was 2015 and one of my biggest learnings was that users clicked emojis!
I was dealing with real VC budgets finally and was allocating how we spend about 50k/mo in advertising. (I was naive these folks didn't have much runway to spend that) Their social networks blossomed, but they weren't monetizing the app itself initially so they shut their doors.
I remember filing for unemployment and filling out an application to a Master's program the same day. I searched for less expensive programs and while I considered big name schools I was not about to drop 60k on a Master's after getting an undergrad in 5 yrs.
I found @RegisUniversity's Data Science Master's and got started. A friend of mine worked for a company in cannabis marketing and I had known the CEO in the Canna-Startup scene for a few years. When she said she needed a Data Scientist it was the perfect opportunity.
In hindsight it was nuts, but I was working 60+ hour weeks for this company, understanding the available market, predicting our customers' churn, recruiting new advertisers, and ultimately selling them when we transitioned to being a SaaS company.
This hands on experience coupled with education helped me learn the ropes a lot quicker.
(Fyi cannabis companies can't use most digital marketing like FB, IG, Twitter ads)
It was like drinking through a firehose, but I was predicting how marketing campaigns would perform YoY.
(Fyi cannabis companies can't use most digital marketing like FB, IG, Twitter ads)
It was like drinking through a firehose, but I was predicting how marketing campaigns would perform YoY.
I analyzed our customer's activity on our site and helped our marketing folks optimize drip campaigns to our customers. Obvs we saw an influx of sales in Feb & Mar for the 4/20 surge. The one time we push publishers to offer sales we saw > 100k MORE than previous yrs.
I did that for a year then took 6 months to focus on school. I was starting to hit the hard courses like Machine Learning and Experiment Design. I then got asked on twitter to do ML for a drone company! The use case was to reduce deaths from gun violence & police brutality.
I was part of writing the paper on what neural network and armored technology the drone uses. It was my first peer-reviewed paper and accepted at the @IEEEorg Global Humanitarian Tech Conference.
https://ieeexplore.ieee.org/document/8601597
https://ieeexplore.ieee.org/document/8601597
My work specifically meant training neural networks on TONS of sensor data to identify different kinds of firearms from each other. We tested radio frequency sensors (like TSA scanners) on .45mm and 9mm weapons and magazines.
My model was surprisingly good, but even 90%+ models in this case are
A) under high scrutiny - 90% confident isn't very confident when it comes to human life
B) public record and should be substituted for interpretable models
A) under high scrutiny - 90% confident isn't very confident when it comes to human life
B) public record and should be substituted for interpretable models
After that role I finished grad school December '18 and started working for Mindbody (where I currently am) in March.
Right now I'm working on projects like predicting our app downloads, consumer segments, and consumer churn!
Right now I'm working on projects like predicting our app downloads, consumer segments, and consumer churn!
If you're still with me I've got some good tips for those interested in Data Science.
1- You should learn about Ethics before anything else. https://hub.packtpub.com/machine-learning-ethics-what-you-need-to-know-and-what-you-can-do/
1- You should learn about Ethics before anything else. https://hub.packtpub.com/machine-learning-ethics-what-you-need-to-know-and-what-you-can-do/
2- Try your hand at Python
@dataquestio has some great free and interactive courses on using Python (and R!) for Data Science https://www.dataquest.io/directory/
@dataquestio has some great free and interactive courses on using Python (and R!) for Data Science https://www.dataquest.io/directory/
3- Find Data you're interested in
@datadotworld has plenty of user-uploaded data you can sift through. I've played with Dog intelligence data, crime statistics and more. https://data.world/
@datadotworld has plenty of user-uploaded data you can sift through. I've played with Dog intelligence data, crime statistics and more. https://data.world/
4- Learn about good Visualizations.
Fun Ones: @puddingviz
How to make yours better: @storywithdata https://pudding.cool/
Fun Ones: @puddingviz
How to make yours better: @storywithdata https://pudding.cool/
5- Learn SQL
SQL is major when it comes to accessing your company's data. One of my fave FREE resources is the @ModeAnalytics SQL tutorial suggested to me when I interviewed a Facebook. https://mode.com/sql-tutorial/
SQL is major when it comes to accessing your company's data. One of my fave FREE resources is the @ModeAnalytics SQL tutorial suggested to me when I interviewed a Facebook. https://mode.com/sql-tutorial/
6- There's more than just Analytics
Did you know there are Tableau developer jobs making over $100k and focus on the DESIGN of displaying charts and graphs?
https://www.guru99.com/tableau-tutorial.html
Did you know there are Tableau developer jobs making over $100k and focus on the DESIGN of displaying charts and graphs?
https://www.guru99.com/tableau-tutorial.html
7- If you like math, there's money to be made.
Let yourself go down rabbit holes while learning ML. If it turns out you really like computer vision or natural language processing there's LOTS of money to be made.
Let yourself go down rabbit holes while learning ML. If it turns out you really like computer vision or natural language processing there's LOTS of money to be made.
8- Learn what a Neural Net is about
Data Science in general and really ML/AI scares people off because of the math involved. If you see how cool the end will be, the means make sense. There are also great resources now for making it not intimidating.
Data Science in general and really ML/AI scares people off because of the math involved. If you see how cool the end will be, the means make sense. There are also great resources now for making it not intimidating.
9- Have fun with it!
ML/AI is "cutting edge" but it's also really fun to learn about. https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle®Dataset=reg-plane&learningRate=0.03®ularizationRate=0&noise=0&networkShape=4,2&seed=0.40199&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false
ML/AI is "cutting edge" but it's also really fun to learn about. https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle®Dataset=reg-plane&learningRate=0.03®ularizationRate=0&noise=0&networkShape=4,2&seed=0.40199&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false
10 - DM me!
If you want to get a Data Science job I'm still taking on 2 new mentees this winter. https://mentors.sharpestminds.com/bKeZkGdXSRQyuXSQ7
DASSIT y'all! Thanks for sticking w/ this long ass thread.
Tomorrow I'm going into Ethics. #30DaysofThreads
If you want to get a Data Science job I'm still taking on 2 new mentees this winter. https://mentors.sharpestminds.com/bKeZkGdXSRQyuXSQ7
DASSIT y'all! Thanks for sticking w/ this long ass thread.
Tomorrow I'm going into Ethics. #30DaysofThreads
