Discover and read the best of Twitter Threads about #TimeSeries
Most recents (24)
Are Transformers useless for time series forecasting? NO! We showcase how Transformers achieve SOTA results, beating DLinear models.
Here's a blog post introducing the Autoformer model and how effective it can be at #timeseries #forecasting: huggingface.co/blog/autoformer
Here's a blog post introducing the Autoformer model and how effective it can be at #timeseries #forecasting: huggingface.co/blog/autoformer
The paper "Are Transformers Effective for Time Series Forecasting?" arxiv.org/abs/2205.13504, claims that a simple linear model DLinear outperforms Transformer based models by a large margin.
Why does that happen in their case? Firstly, they compare a linear model with massive models, and they even acknowledge it. Secondly, they compare their univariate model with multivariate versions of the Transformers, which on the small datasets gives poor performance.
1/ 🧪🔬 Ever wondered if scientific research can be predicted? Our latest paper "What's next? Forecasting scientific research trends" tackles this question. Check out the details here 👉 arXiv:2305.04133v1 [thread] 🧵
2/ 🌍 The world of science is constantly evolving. But what if we could predict these trends, especially in life sciences? That's what our work aims to do. Imagine the possibilities!
3/ 📚 We've mined historical publications, research/review articles, and patents. The result? Models that can foresee scientific trends 5 years in advance. Time travel, anyone? #AI #DataScience #timeseries #trends
Learning data science on your own is tough...
...(ahem, it took me 6 years)
So here's some help.
5 Free Books to Cut Your Time In HALF.
Let's go! 🧵
#datascience #rstats #R
...(ahem, it took me 6 years)
So here's some help.
5 Free Books to Cut Your Time In HALF.
Let's go! 🧵
#datascience #rstats #R
1. Mastering #Spark with #R
This book solves an important problem- what happens when your data gets too big?
For example, analyzing 100,000,000 time series.
You can do it in R with the tools covered in this book.
Website: therinspark.com
This book solves an important problem- what happens when your data gets too big?
For example, analyzing 100,000,000 time series.
You can do it in R with the tools covered in this book.
Website: therinspark.com
2. Geocomputation with #R
Interested in #Geospatial Analysis?
This book is my go-to resource for all things geospatial.
This book covers:
-Making Maps
-Working with Spatial Data
-Applications (Transportation, Geomarketing)
Website: r.geocompx.org
Interested in #Geospatial Analysis?
This book is my go-to resource for all things geospatial.
This book covers:
-Making Maps
-Working with Spatial Data
-Applications (Transportation, Geomarketing)
Website: r.geocompx.org
99% of data scientists CANNOT do #timeseries and #automation together.
I know this because almost all of my advanced students ask for help in these 2 areas.
So I made a FREE training.
Here's why + how to access it for free... 🧵
#datascience #rstats
I know this because almost all of my advanced students ask for help in these 2 areas.
So I made a FREE training.
Here's why + how to access it for free... 🧵
#datascience #rstats
What if you could use any #Python or #R library to do it?
Which would you pick?
Here's what I'd do and why.
Which would you pick?
Here's what I'd do and why.
1. Modeltime in R for Time Series
It's not a question. It's a fact.
Modeltime is the best forecasting library in R or Python ecosystems.
Sure the bar is high:
- forecast
- fable / tsibble
- statsmodels
- sktime
- pmdarima
- darts
- +20 more
Modeltime is the best.
It's not a question. It's a fact.
Modeltime is the best forecasting library in R or Python ecosystems.
Sure the bar is high:
- forecast
- fable / tsibble
- statsmodels
- sktime
- pmdarima
- darts
- +20 more
Modeltime is the best.
🎉 We are thrilled to announce the release of the latest version of mlforecast a #Python library for Scalable #machinelearning 🤖 for #timeseries #forecasting
🚀 This version comes with exciting new features that are sure to make forecasting even more efficient and accurate
🧵
🚀 This version comes with exciting new features that are sure to make forecasting even more efficient and accurate
🧵
🔮 Conformal Prediction: We've added the ability to generate probabilistic forecasts using conformal prediction. You can choose the levels of your intervals and the number of conformity scores to use.
⚡️ In addition, mlforecast includes the capability to fit multiple time series with global models. That means you don’t need to calculate sequential conformal intervals for each series and period. I.e., conformal prediction for time series just got a lot faster and easier. 🤯
🎉 We are very excited to release the new features of NeuralForecast! 🥳🚀
With this release time series forecasting with neural models is even more accessible and powerful. Here are some of the highlights
🧵
#timeseries #python #deeplearning #forecasting
With this release time series forecasting with neural models is even more accessible and powerful. Here are some of the highlights
🧵
#timeseries #python #deeplearning #forecasting
📈 New Models:
- Temporal Convolution Network
- AutoNBEATSx
- AutoTFT (Transformers)
- Temporal Convolution Network
- AutoNBEATSx
- AutoTFT (Transformers)
🧮 Recurrent models (RNN, LSTM, GRU, DilatedRNN) can now take static, historical, and future exogenous variables.
💥14 herramientas secretas impulsadas por #RStats para ahorrar tiempo y esfuerzo en tus proyectos de datos (¡No te lo pierdas!):👀
1️⃣ ¡Edita tus datos de forma interactiva (y guarda el código)! 👀
📦 'editData' es un complemento de RStudio para editar un data.frame o un tibble de forma interactiva
🔗 buff.ly/3U5Tgjy
#DataScience #DataVisualization #dataviz #stats #analytics #RStats #Analytics
📦 'editData' es un complemento de RStudio para editar un data.frame o un tibble de forma interactiva
🔗 buff.ly/3U5Tgjy
#DataScience #DataVisualization #dataviz #stats #analytics #RStats #Analytics
2️⃣ ¡Crea gráficos #ggplot de forma interactiva!🚀
📦esquisse es otro de mis addins favoritos de #rstudio
✅ exporta el gráfico o recupera el código para reproducir el gráfico
🔗 buff.ly/3mxLHSo
#DataScience #DataVisualization #dataviz #stats #analytics #RStats #Analytics
📦esquisse es otro de mis addins favoritos de #rstudio
✅ exporta el gráfico o recupera el código para reproducir el gráfico
🔗 buff.ly/3mxLHSo
#DataScience #DataVisualization #dataviz #stats #analytics #RStats #Analytics
What's the difference between Time Series Analysis and Forecasting? 🤔
🧵 THREAD 🧵
#machinelearning #python #datascience #ai
🧵 THREAD 🧵
#machinelearning #python #datascience #ai
Time Series Analysis and Time Series Forecasting are both methods used to analyze and make predictions about data collected over time, but they have different focuses and applications.
▶️ Time Series Analysis is the process of understanding the underlying structure and patterns in a dataset.
It is used to identify trends, patterns, and relationships in the data, as well as to examine the behavior of a variable over time.
It is used to identify trends, patterns, and relationships in the data, as well as to examine the behavior of a variable over time.
Let's talk about nomenclature in Time Series Forecasting ✏️
🧵 THREAD 🧵
#MachineLearning #DataScience #Python
🧵 THREAD 🧵
#MachineLearning #DataScience #Python
Time series data is a sequence of observations taken at specific time intervals. To understand it, we need to establish some key terms.
The current time is represented as t.
The observation made at the current time is represented as obs(t).
The observation made at the current time is represented as obs(t).
Want to see the power of Time Series forecasting in action? 📈
Let's take a look at how the retail giant Walmart uses it to improve its operations. 🤯
🧵 THREAD🧵
#timeseries #forecasting #retail #example
Let's take a look at how the retail giant Walmart uses it to improve its operations. 🤯
🧵 THREAD🧵
#timeseries #forecasting #retail #example
Walmart has been using time series forecasting for decades to predict sales and optimise inventory levels.
By analysing historical sales data, weather patterns, and other factors, they're able to make more accurate predictions about future demand for products.
By analysing historical sales data, weather patterns, and other factors, they're able to make more accurate predictions about future demand for products.
This allows them to ensure they have the right products in stock at the right time, reducing stockouts and increasing customer satisfaction.
How can you transform your Time Series data to make it stationary? 🤔
Check below 7️⃣ ways of achieving this! 👇
🧵 THREAD🧵
#python #datascience #machinelearning #ai
Check below 7️⃣ ways of achieving this! 👇
🧵 THREAD🧵
#python #datascience #machinelearning #ai
1️⃣ Differencing:
This method involves subtracting the previous value from the current value.
This can help to remove any trend or seasonality in the data and make it stationary.
This method involves subtracting the previous value from the current value.
This can help to remove any trend or seasonality in the data and make it stationary.
2️⃣ Moving Average:
This method involves taking the average of the time series over a certain period, such as a month or a year.
This can help to smooth out any fluctuations in the data and make it more stationary.
This method involves taking the average of the time series over a certain period, such as a month or a year.
This can help to smooth out any fluctuations in the data and make it more stationary.
Don't forecast future values in Time Series using the traditional way!! ⛔
Discover the Rolling forecasting! 🤯
#python #machinelearning #ml #ai #datascience
Discover the Rolling forecasting! 🤯
#python #machinelearning #ml #ai #datascience
Yesterday we discussed the first way of forecasting with your Time Series model:
1️⃣ The traditional way or multi-step forecast
Today is time for the second (and better) way:
2️⃣ Rolling forecast
1️⃣ The traditional way or multi-step forecast
Today is time for the second (and better) way:
2️⃣ Rolling forecast
2️⃣ Rolling forecast
As mentioned yesterday, this consists of training the model every day, with all the available data until the present day.
Then we forecast for tomorrow.
As mentioned yesterday, this consists of training the model every day, with all the available data until the present day.
Then we forecast for tomorrow.
Be careful how you forecast the future in Time Series! 🤔
There are two main ways:
1️⃣ The traditional way or multi-step forecast
2️⃣ Rolling forecast
Let's see how they differ 👇
🧵 Thread🧵
There are two main ways:
1️⃣ The traditional way or multi-step forecast
2️⃣ Rolling forecast
Let's see how they differ 👇
🧵 Thread🧵
🔎 Let's consider the Apple stock price for this example.
⚠️ However, this is applicable to any Time Series data!
⚠️ However, this is applicable to any Time Series data!
1️⃣ The traditional way or multi-step forecast
This consists of training the model once, with all the available data.
Then we forecast for several days in the future.
This consists of training the model once, with all the available data.
Then we forecast for several days in the future.
Many courses teach #DataScience libraries such as pandas, matplotlib and seaborn, but @TedPetrou’s self-paced online course “Master Data Analysis with #Python” teaches BEST PRACTICES in using the libraries to help you become an expert!
⭐⭐
Start here now: dunderdata.com/a/16370/bofLf2…
⭐⭐
Start here now: dunderdata.com/a/16370/bofLf2…
532-page #Python #coding book by @TedPetrou >> "Pandas Cookbook: Recipes for Scientific Computing, #TimeSeries Analysis, and Data Visualization" at amzn.to/3KOQP1v
—————
#100DaysOfCode #DataScience #DataScientists #DataViz #MachineLearning #ComputationalScience
—————
#100DaysOfCode #DataScience #DataScientists #DataViz #MachineLearning #ComputationalScience
3/
Explore these self-paced online courses that help you master the tools of #DataScience:
1) Master #MachineLearning with #Python: dunderdata.com/a/16371/bofLf2…
2) Build a Data Analysis Library from Scratch in Python: dunderdata.com/a/16373/bofLf2…
3) All-Access Pass: dunderdata.com/a/16369/bofLf2…
Explore these self-paced online courses that help you master the tools of #DataScience:
1) Master #MachineLearning with #Python: dunderdata.com/a/16371/bofLf2…
2) Build a Data Analysis Library from Scratch in Python: dunderdata.com/a/16373/bofLf2…
3) All-Access Pass: dunderdata.com/a/16369/bofLf2…
Thanks ICP for publishing this essay on #Ritamic Decision Policy. Into year-3 of this series, this post summarizes manthan, research and study of the Vedic origins of sustainable decision policy and strategy, with contemporary examples all can relate to.
Post focuses on sustainable policy to make not one, but a series of interconnected decisions over time.
Can ideas of Vedanta be applied here?
short ans: Y.
Content is of interest to those making gov or private policy, designers, startups, #Ganita/stem students, young parents.
Can ideas of Vedanta be applied here?
short ans: Y.
Content is of interest to those making gov or private policy, designers, startups, #Ganita/stem students, young parents.
The post is divided into 7 sections with links to each at the top. Those who simply want a easy-to-remember idea of Ritamic decision policy can go to the examples- read how Air India altered its DEL-SFO route in harmony with Ritam. flightradar24.com/blog/wp-conten…
Hierarchical forecasting is a unique time series problem with several challenges because of time series dependencies.
An example is Walmart Store Sales from the M5 Competition.
📈 Walmart has stores in each state.
📉 Each state has multiple stores.
📈 Each store has multiple departments.
📉 Each department has categories of items.
📈 Each category of items has 100s or 1000s of items.
📈 Walmart has stores in each state.
📉 Each state has multiple stores.
📈 Each store has multiple departments.
📉 Each department has categories of items.
📈 Each category of items has 100s or 1000s of items.
Did you know about Facebook Prophet?
It's an open-source tool for forecasting Time Series data, and it's super easy to use!
Find out more about it!
🧵 👇
#DataScience #MachineLearning #Python #Facebook #AI
It's an open-source tool for forecasting Time Series data, and it's super easy to use!
Find out more about it!
🧵 👇
#DataScience #MachineLearning #Python #Facebook #AI
Facebook Prophet can handle missing data and changes in trends, which are common in time series data.
One of the best things about Facebook Prophet is its flexibility.
You can fine-tune your forecasts with a variety of parameters.
You can fine-tune your forecasts with a variety of parameters.
Time Series can be used for a variety of applications.
Here you have 5️⃣ applications or possible projects where Time Series Forecasting is key!
#DataScience #MachineLearning #Python #AI
Here you have 5️⃣ applications or possible projects where Time Series Forecasting is key!
#DataScience #MachineLearning #Python #AI
1️⃣ Forecasting demand for a product or service:
This could involve collecting historical data on the quantity of a product or service sold, and using time series techniques to forecast future demand.
This could involve collecting historical data on the quantity of a product or service sold, and using time series techniques to forecast future demand.
2️⃣ Predicting stock prices:
This could involve collecting historical data on the closing price of a stock, and using a time series model to make predictions about future prices.
This could involve collecting historical data on the closing price of a stock, and using a time series model to make predictions about future prices.
Conformal prediction for #timeseries #forecasting is the hottest 🔥🔥🔥🔥 🚀🚀🚀🚀 area of research and applications right now.
The whole story started in summer 2021 when researchers from Georgia Tech published the first ever paper applying conformal prediction to time series.
The whole story started in summer 2021 when researchers from Georgia Tech published the first ever paper applying conformal prediction to time series.
and presenting it at the first event at major conference totally dedicated to conformal prediction (ICML DFUQ 2021 organised by @ml_angelopoulos and @stats_stephen
Roll forward just 1.5 yeast - there are multiple papers published on the subject, several #opensoure libraries like MAPIE implemented the original first ever paper from Georgia Tech (method called EnbPI).
What is the main difference between ARCH and GARCH models? 🤔
Find it out here!
🧵 👇
#python #machinelearning #timeseries #datascience
Find it out here!
🧵 👇
#python #machinelearning #timeseries #datascience
🔎 GARCH stands for "Generalized Autoregressive Conditional Heteroskedasticity".
It is a type of time-series model that is commonly used in finance to model the volatility of asset returns.
It is a generalisation of the ARCH model.
But why a generalisation? 👇 🤔
It is a type of time-series model that is commonly used in finance to model the volatility of asset returns.
It is a generalisation of the ARCH model.
But why a generalisation? 👇 🤔
It extends the capabilities of ARCH models. ⚡
It allows for the inclusion of additional variables and terms in the model.
This can improve its accuracy and ability to capture the dynamics of the time series. 🤩
Let's see how 👇
It allows for the inclusion of additional variables and terms in the model.
This can improve its accuracy and ability to capture the dynamics of the time series. 🤩
Let's see how 👇
5️⃣ YouTube playlists / videos to learn Time Series! ▶️
Check them out!
🧵 👇👇
#Python #MachineLearning #DataScience
Check them out!
🧵 👇👇
#Python #MachineLearning #DataScience
1️⃣ "Time Series Analysis" by ritvikmath
My personal top 1 recommendation for learning Time Series.
A great combination of theory and code 👌
youtube.com/playlist?list=…
My personal top 1 recommendation for learning Time Series.
A great combination of theory and code 👌
youtube.com/playlist?list=…
2️⃣ "Time Series" by Aric LaBarr
👍 Good variety of models. Short videos.
👎 Purely theoretical, no code.
youtube.com/playlist?list=…
👍 Good variety of models. Short videos.
👎 Purely theoretical, no code.
youtube.com/playlist?list=…
Machine Learning and Deep Learning are key skills for a Data Scientist! 🔑 But also for Time Series! 📈
TOP 5️⃣ COURSES to learn about it 👇
👇👇👇👇👇
#python #datascience #ai
TOP 5️⃣ COURSES to learn about it 👇
👇👇👇👇👇
#python #datascience #ai
1️⃣ Start by learning the basics of Machine Learning with this fantastic course.
• Learn both supervised and unsupervised algorithms
• Get introduced to Neural Networks
• Find out about XGBoost
coursera.org/specialization…
• Learn both supervised and unsupervised algorithms
• Get introduced to Neural Networks
• Find out about XGBoost
coursera.org/specialization…
2️⃣ Learn about Deep Learning:
• Neural Networks
• Convolutional Neural Networks (CNN)
• Sequence models (very useful for Time Series!)
coursera.org/specialization…
• Neural Networks
• Convolutional Neural Networks (CNN)
• Sequence models (very useful for Time Series!)
coursera.org/specialization…
What is the difference between seasonality and cycle in Time Series? 🤔
🧵 👇
#Python #DataScience #MachineLearning
🧵 👇
#Python #DataScience #MachineLearning
🔴Seasonality refers to regular patterns that occur at a specific frequency, often at a yearly, monthly or weekly interval.
For example:
• Retail sales tend to increase during the holiday season
• Electricity consumption tends to be higher in the summer months when people use air conditioning more often
• Retail sales tend to increase during the holiday season
• Electricity consumption tends to be higher in the summer months when people use air conditioning more often
ARCH can improve your ARIMA Time Series forecast!
Learn how 👇👇
A thread 🧵
#Python #MachineLearning #DataScience
Learn how 👇👇
A thread 🧵
#Python #MachineLearning #DataScience
We've seen that ARCH is a model to forecast the variance of a time series.
It is frequently used in situations in which there may be short periods of increased variation or volatility.
It is frequently used in situations in which there may be short periods of increased variation or volatility.
They were created for finance and econometric problems. 💸
But...
But...
