90% of data scientists overlook how to design A/B Testing experiments.

4 tips for better experiments: 🧵

#DataScience #ABTesting Tip 1: Include a pre-test

Pretest data is unaffected data before the actual A/B test or Time-based Experiment.

Pre-test is a secret used by Booking(dot)com in their CUPED A/B Test method for reducing variance (and improving decision-making from A/B Test results).

Both Bayesian and Frequentist approaches to A/B testing have strengths (and weaknesses).

Here's a quick selection guide with 4 Pros/Cons. 🧵

#Bayesian #Frequentist #MachineLearning #ABTesting 💡 4 Reasons for the #Frequentist Approach for A/B testing

1. Fixed Sample Size: Requires pre-determination of sample size. Ideal when sample size cannot change once the test begins.

Stop using frequentist approaches for A/B Testing.

Use Bayesian instead.

Bayesian has 5 key advantages: 🧵

#DataScience #Bayesian #Rstats #Python 1. Intuitive Interpretation:

Bayesian methods provide results in terms of probabilities.

Bayesian probabilities are more intuitive to understand AND more accurate compared to t-test or linear regression p-values.

12 mistakes that Data Scientists (and even statisticians) make:🧵

#DataScience #Statistics #DataAnalysis #CommonMistakes #CriticalThinking #DataIntegrity Even seasoned data professionals can fall into traps.

Here are 12 common mistakes and misconceptions in statistics and data science:

1. Correlation vs Causation 🔄

Mistaking correlation for causation is a classic error! Remember, correlation does not imply causation. 🚫

Can ChatGPT be used for Time Series?

A thread with #R code.

#rstats I've been using ChatGPT a lot more. But one question I had is whether or not it could be used for Time Series.

In this thread, we'll:

1. Show that chatgpt can write time series code in R
2. Provide code examples
3. Show the app that ChatGPT built for me 👇

90% of data scientists struggle with Price Elasticity and Optimization.

Why?

Outliers.

This is how to save your company. (And your career) 🧵

#datascience #stats #BusinessAnalytics Demand is not static. It's constantly changing. And this costs Data Scientists companies and their careers.

But there's a new technique that can help.

Quantile GAMs.

90% of data scientists struggle with price optimization.

Demand patterns are complex.

I have good news. 🧵

#datascience #PriceElasticity #Optimization #Python #Rstats Demand patterns are complex.

The competitive landscape is ever-changing.

And prices are elastic, which is costing these companies (and their careers).

12 mistakes that Data Scientists (and even statisticians) make:🧵

#DataScience #Statistics #DataAnalysis #CommonMistakes #CriticalThinking #DataIntegrityEven Even seasoned data professionals can fall into traps.

Here are 12 common mistakes and misconceptions in statistics and data science:

1. Correlation vs Causation 🔄

Mistaking correlation for causation is a classic error! Remember, correlation does not imply causation. 🚫

My new timetk for python package just got an upgrade!

Time series plotting.

Here's the details. 🧵

#datascience #timeseries #rstats #python If you're familiar with Time Series in #R then you've probably seen me use my Timetk in R package (2,000,000+ downloads).

Time series plotting is one of its best features.

Link: business-science.github.io/timetk/article…

Introducing our new time series package in Python.

And how to learn Time Series Analysis in R and Python for free.

This is how. 🧵

#DataScience #Rstats #Python over the past 5 years, I've developed dozens of open-source time series packages in R.

Timetk for R is my most popular time series package with over 2,000,000 downloads.

And time series R fans love anomalize, modeltime, and the 5+ extensions my students and I have developed.

When Clustering in Python for Customer Segmentation...

PCA is a secret weapon in the struggle to scale beyond 50,0000 customers.

Here's why. 🧵

#datascience #analytics #python A common business case is to apply K-means to 100,000 or even 1,000,000 customers.

K-means takes several hours because as databases have increased in size, algorithms like K-Means and DBSCAN slow down.

I've spent the past 4 weeks researching ways to speed it up.

Python's Scikit Learn has 12 clustering algorithms.

With all these options, why do 90% of data scientists still struggle with clustering and customer segmentation?

#datascience #analytics #python #rstats 👉 Too many algorithms

More algorithms makes it challenging to select the right approach.

Fortunately, all of these options are not created equal.

I prefer K-means, HDBCAN, but I like them for different reasons.

Outliers and scalability can influence which I use and when.

Clustering is a superpower.

Learn it and you're an unstoppable force.

Use these 5 powerful tricks to master clustering in 10 minutes:

#datascience #python #rstats 90% of data scientists struggle with clustering.

As databases increase in size, methods like K-means or DBSCAN can fail or become slow.

But I have 5 power-tricks to help you cluster like a PRO:

I put together these 5 time series tutorial to help you learn more than a college degree will teach you.

Learn in 7 minutes what took me 7 years.

#datascience #rstats 1. (Start Here) Time Series Forecasting Foundations

Use this to immediately 10X your companies forecasting capability.

This is the foundation of forecasting...

...with my Modeltime Forecasting framework.

Article: https://t.co/zFXL8yYJoqbuff.ly/3RJmVio

ChatGPT was a struggle for me at first.

And, it actually took me months to confidently crush projects with it.

So I put together 5 tips that I wish someone would have told me.

Here you go.

#datascience #chatgpt 1. Start small:

A big mistake is trying to do something 'monumental'

Instead, do small things that add up to big things.

It's 2023 and traditional data science is broken.

There's a new type of Data Professional forming.

Here are the 2023 #datascience #trends I'm seeing... 🧵

#datascientist #career In the past, companies were loading up on bulky data science teams with:

3 Data Scientists ($129,680/yr) 👨‍🔬👩‍🔬👨‍🔬
2 Data Engineers ($129,244/yr) 👨‍🔧👩‍🔧
2 Machine Learning Engineers ($138,244/yr) 👨‍🍳👩‍🍳
2 Data Analysts ($83,924/yr) 👨‍💼👩‍💼
1 Project Manager ($140,500/yr) 🕵️

Did you know that R has a killer forecasting ecosystem?

Here's everything you need to know in under 2 minutes. 🧵

#rstats #datascience #timeseries R's forecasting ecosystem is ABSOLUTELY a thing of beauty.

You get:

1. Time Series Analysis
2. Forecasting
3. Ensembles
4. Resampling & Backtesting
5. AutoML
6. Deep Learning

(full disclosure - I created all the R packages I'm about to cover)

I've been experimenting with #chatgpt for #datascience for 16 weeks.

And I now have a process I'm happy with.

Here are the details. 🧵

#datascience #rstats Using ChatGPT for data science has been a MASSIVE learning curve.

I began using it for complex workflows.

And I FAILED miserably.

It took me 5-years to feel confident in data science.

True story.🧵

#datascience #rstats This is coming from a person that has created two R packages that combine for 1.5 Million downloads.

Has trained elite data scientists at Apple, Walmart, Google.

And has built a career teaching students how to become data scientists.

Why did it take so long?

There are over 2,000 AI tools that have hit the market over the last 365 days.

So I condensed them into the best.

Here are the TOP 15 AI TOOLS for Data Scientists. 🧵

#datascience #rstats #python #career #ai It's hard not to get excited about #AI. The potential is insane. It's also scary.

And the worst thing you can do for your career is ignore AI.

I mean, there are literally 2,000 new tools that have hit the market in 365 days. So where do you start?

I want to help.

3 battle-tested skills that every data scientist should have.

(and how to apply them to a job interview)🧵

#datascience #skills #rstats #python People don’t realize this but I was a data science consultant and corporate trainer...

That was long before I was a “teacher” and a “6-figure data science mentor”.

That’s where I learned these skills through battle-testing.

And my clients were my test subjects. 🧪🧑‍🔬