Tweet

How to get URL link on X (Twitter) App

On the Twitter thread, click on or icon on the bottom
Click again on or Share Via icon
Click on Copy Link to Tweet
Paste it above and click "Unroll Thread"!
More info at Twitter Help

Levi

@levikul09

Oct 21 • 9 tweets • 3 min read Twitter logo

Read on Twitter

Scrolly

How to describe distribution?

Skewness 🧵

Skewness is a measure of the asymmetry of a distribution.

Distribution is symmetric if it looks the same to the left and right of the center point.

Skewness differentiates extreme values in one versus the other tail.

1/8

The formula to get the skewness:

2/8

Left skew / Negative skew:

Negative values for the skewness indicate data that are skewed left - the left tail is long relative to the right tail.

The mean of a left-skewed distribution is almost always less than its median.

3/8

Right skew / Positive skew:

Positive values for the skewness indicate data that are skewed right - the right tail is long relative to the left tail.

Here the mean is almost always greater than the median.

Because extreme values affect the mean more than the median.

4/8

Zero skew:

When a distribution has zero skew, it is symmetrical. Its left and right sides are mirror images.

The skewness for a normal distribution is zero.

The mean = median

5/8

Pandas has a built-in method to calculate the skewness of the data.

Since the value is negative, the data is skewed to the left - the left tail is slightly longer than the right tail.

6/8

https://twitter.com/levikul09/status/1715671552254189727

That's it for today.

I hope you've found this thread helpful.

Follow me @levikul09 for more.

Like/Retweet the first tweet below for support, Thanks 😉

7/8

https://twitter.com/levikul09/status/1715671552254189727

If you haven't already, join our newsletter DSBoost.

We share:

• Interviews

• Podcast notes

• Learning resources

• Interesting collections of content

8/8dsboost.dev

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @levikul09

Levi

@levikul09

Oct 18

Backpropagation is fundamental in ANNs

But how does it work?

I will explain it now visually.

1/9

Consider this:

We have a model that you try to fit to the points. This model is a result of a NN.

It is clear that this NN provides a bad fit, so we need to adjust the parameters.

How can we do that?

Let's see step by step.

2/9

1. We calculate the errors.

Comparing the observed values with the predicted values will give us the errors.

If we sum the errors we can plot them 🔽

(Note: this is not a thread on error metrics you can think of any of them.)

3/9

Read 9 tweets

Levi

@levikul09

Oct 15

6 Statistical and Machine Learning pitfalls.

Avoid these traps to be a better data person.

1/9

1️⃣ Correlation = Causation

They are related, it is crucial to understand that correlation does not imply causation!

We cannot measure causation statistically!

Resist the temptation to build a causal narrative around correlated variables.

2/9

2️⃣ P-hacking

Statistically significant results do not always imply real-life significance.

Studies can manipulate or selectively analyze data in order to obtain statistically significant results.

It is important to follow transparent and rigorous research practices.

3/9

Read 9 tweets

Levi

@levikul09

Oct 14

Regression to the mean is often misunderstood.

It refers to extreme events followed by average outcomes.

But it's not just randomness.

Let me explain

🧵

Do you know what the "Madden Curse" is?

Madden is an American football video game.

The game usually features the best players from the NFL on the cover.

For most players being on the cover means the next season will be worse.

Why?

1/10

Of course, there is no curse at all.

It is because of the regression to the mean.

When variables are significantly higher or lower than average on the first measurement, they move closer to the average on the second measurement.

2/10

Read 11 tweets

Levi

@levikul09

Oct 13

Weights and Biases are the engines in Neural Networks.

I will explain how they work.

1/8

When data is flowing between different neurons or layers, it is not just going from A to B.

Different transformations happen to them.

These transformations are described with Weights and Biases.

Let's discuss each 🔽

2/8

1️⃣ Weight

Weights determine how important each factor is in the overall prediction.

This value will determine the influence input data has on the output product.

They work similarly as in weighted means: The input is multiplied by the weights.

3/8

Read 8 tweets

Levi

@levikul09

Oct 12

Free Datasets from tech giants.

Microsoft, Google, IBM, Amazon...

Their valuable data is just a few clicks away.

Find the links below 👇

1️⃣ Registry of Open Data on AWS

This registry exists to help people discover and share datasets that are available via AWS resources.

registry.opendata.aws

2️⃣ Data Asset eXchange by IBM

developer.ibm.com/exchanges/data/

Read 7 tweets

Levi

@levikul09

Oct 11

When you start learning Statistics you may feel there are a million things to memorize.

You're wrong.

What you need is 5-6 basic concepts. Everything else is built around them.

Here are 5+1 Statistical principles to begin with:

🧵

You will never stop learning if you enjoy the journey.

I partnered with @brilliantorg for this reason. They introduced the joy of studying Data Science.

The best part?

They cover all these topics in their intuitive lessons.

Check the link in the last tweet and start learning!

1. Central tendency

Mean, mode, and median are measures that offer quick insights into the 'center' of the data.

It helps to see what a "normal" data point looks like in a group.

We hear about these measures a lot, so it's good to know them well.