Nihilist Data Scientist

Statistically insignificant observations on data science. Submissions: omnia trahe - nihil inveniet - morietur solus

[Nihilist Data Scientist @nihilist_ds]

The learning may be deep, but the understanding is shallow. #DataScience #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

Don’t worry if you fail. It’s usually not even possible to measure the variables necessary to answer the question you’re studying. #rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

Anything worth doing is worth doing with terrible, incomplete data.

#rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

A preoccupation with which tools are popular is an excellent way to distract yourself from the sad truth that everything is dying and no one will remember anything you say or do. #rstats #pydata #datascience

[Valeriy M., PhD, MBA, CQF @predict_addict]

R is dying, if you don't believe me look at THE main conformal prediction library.

When writing my book 'Practical Guide to Applied Conformal Prediction: Learn and apply the best uncertainty frameworks to your industry applications' https://t.co/5I9OJYkNwl I have run a poll asking data scientists and others whether they use Python or R for their work.

The results came overwhelmingly in favour of Python, R was less than 20% and mostly academics.

The same trend is happening in many areas also in time series and forecasting.

I have been using R since 2012 and it is sad, but the trends are clear and irreversible.

You can still find R section on Awesome Conformal Prediction but unlike Python section it is not growing as there are hardly any new libraries.

Python vs R for time-series forecasting https://t.co/mUJ0CPfS4g

The best bet for this library is to port everything that is still missing in Python into most popular libraries like MAPIE and Amazon Fortuna.

#python #datascientists #R

[Nihilist Data Scientist @nihilist_ds]

#datascience in a corporate setting is 80% data cleaning.

The other 20% is doing the analysis to convince the C-suite that you were correct when you told them at the start that the question wouldn't be answerable from the data you have. #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

What it means when I have N tabs in my browser open to StackOverflow git questions:

0: I've written no code today

3: I've made the same mistake I make every 3mo but can never remember how to fix

10: Pure, unadulterated panic

#rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

Most data contains no insights, including the data in this tweet.

#rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

If you're unsure whether you should switch to the #rstats native pipe |> from %>%, I recommend a compromise: alternate between the two on each line.

data |>
select() %>%
filter() |>
group_by() %>%
summarise()

[April King 🌀 @CubicleApril]

I'm starting to think I have what it takes to become a data scientist.

[Nihilist Data Scientist @nihilist_ds]

If you're alone today on Valentine's Day, remember: nothing can impute the missing data in your heart.

#rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

No, the data engineers w/access to the HIPAA protected data can't run a 5 line SQL query to get one number.

They are too busy building the secure data pipeline that will allow you to self-serve. It will be done in 6 months, same as when you asked 6 months ago. #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

If @JosiahParry ever needs data from me, it shall be provided in the format:

())(value)(()

(Just kidding, I never respond to any requests for data because we'll all be dead soon anyway.)

[Nihilist Data Scientist @nihilist_ds]

Remember, a forecast based on a model is a just a guess wearing a top hat.

#datascience #rstats #pydata #statistics

[Nihilist Data Scientist @nihilist_ds]

Humans like to tell themselves that the collection of passing moments in their lives converges in distribution to happiness but all we know for sure is that they converge almost surely in death. #datascience #statistics #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

Your periodic reminder that the vast majority of data sets you will be asked to investigate will be boring and uninformative.

So....don't get excited.

#rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

This is even better, tbh. #datascience #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

Well done, ChatGPT, well done. #datascience #rstats #pydata

[Nihilist Data Scientist @nihilist_ds]

The data pipeline process that's been in production for 10 years that no one dares touch. #rstats #pydata #datascience

[Nihilist Data Scientist @nihilist_ds]

Time series forecasting is like being alive. Futile, but necessary. #rstats #datasciece #pydata https://t.co/g2WrdszoaU

[Nihilist Data Scientist @nihilist_ds]

In practice, the #DataScience method is to start with the Expert method and then back your way into the Novice method from that. #rstats #pydata https://t.co/9huhRQUmj3