EconML
mta
EconML | mta | |
---|---|---|
8 | 3 | |
3,557 | 91 | |
1.5% | - | |
8.5 | 0.0 | |
1 day ago | about 2 years ago | |
Jupyter Notebook | Python | |
GNU General Public License v3.0 or later | - |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
EconML
-
[D] What approach to decide which class is most optimal for recovery?
A good package with many of the tools used for this type of problem as well as pretty good documentation about how it all works is https://github.com/microsoft/EconML
- Getting treatment effects from a random forest
-
Data Science and Marketing
Uplift Modeling (python): CausalML, EconML
-
EconML: A Python Package for ML-Based Heterogeneous Treatment Effects Estimation
Github: https://github.com/microsoft/EconML
-
UpliftML: An uplift modeling library that handles web scale datasets
Many libraries have recently emerged that offer implementations of algorithms for heterogeneous treatment effect estimation (or, CATE estimation). The most well-known examples are Microsoft's EconML (https://github.com/microsoft/EconML) and Uber's CausalML (https://github.com/uber/causalml). Existing libraries require all data to fit in memory, which is often a limitation for industry applications on web scale datasets. Booking.com's new library offers similar functionality on top of Spark, enabling web scale uplift modeling.
-
[q] before/after test
EconML and CausalLift are pretty good python packages that help you build uplift models. scikit-uplift is a decent sklearn style wrapper package that can be helpful as well. One of the drawbacks of these packages is they only allow for the modeling of a single treatment. mr-uplift is a newer package that allows you to model the multiple treatment effects simultaneously. I haven't used it personally, but it does look fairly interesting.
-
[N] Spotify Confidence - open source for analyzing a/b test data
Can't see how this adds to decades of causal inference packages development in stats oriented frameworks like R/Stata/EViews etc and the ongoing effort of porting this to Python. If you want something fancy there's already EconML.
-
What are some applications of Data Science in Digital Marketing?
Uplift Modeling - This is a very powerful technique aimed at discovering the customers who are most likely to respond to your marketing efforts. Some good python libraries for this are EconML and mr-uplift
mta
-
Is Hierarchical Bayesian Modelling used in industry?
Python library of a bunch of attribution models
-
What are some applications of Data Science in Digital Marketing?
Some other marketing topics to be aware of: forecasting - Prophet is an interesting library for this, you'll definitely need some domain knowledge to fit the forecast, it really shouldn't be used to just fit and go otherwise you'll probably end up with some bad results, Media Mix Modeling - FB-Robyn is a library with quite a bit of potential, Multi-Touch Attribution - MTA is a decent python library for this, but you'll have pretty significant data requirements to actually have accurate results, these approaches tend to be pretty susceptible to survivorship/selection bias, survival analysis - Lifelines is a pretty good python package for this, this sort of analysis is useful for determining churn likelihood or time until next purchase.
-
[Marketing Attribution Model for B2B] How to assign revenue based on the lead source?
This is a nice library that implements several multi-touch attribution models beyond the simpler heuristic based ones. One word of caution about these sort of attribution models is the attribution always adds up to 100%. Attribution models typically don't take exogenous factors into account - things that potentially influence whether the customer would have purchased anyway regardless of marketing touchpoints. They also tend to be quite sensitive to selection bias. If you have a touchpoint that requires a customer perform some behavior that can be related to a base level of interest, the model will overweight the attribution of that touchpoint - think things like an abandoned cart remarketing journey. The customer has already shown an inherent interest in the product by placing the product in the cart.
What are some alternatives?
causalml - Uplift modeling and causal inference with machine learning algorithms
Robyn - Robyn is an experimental, AI/ML-powered and open sourced Marketing Mix Modeling (MMM) package from Meta Marketing Science. Our mission is to democratise modeling knowledge, inspire the industry through innovation, reduce human bias in the modeling process & build a strong open source marketing science community.
upliftml - UpliftML: A Python Package for Scalable Uplift Modeling
trimmed_match - This Python library implements Trimmed Match for analyzing randomized paired geo experiments and also implements Trimmed Match Design for designing randomized paired geo experiments.
lightweight_mmm - LightweightMMM 🦇 is a lightweight Bayesian Marketing Mix Modeling (MMM) library that allows users to easily train MMMs and obtain channel attribution information.
causalglm - Interpretable and model-robust causal inference for heterogeneous treatment effects using generalized linear working models with targeted machine-learning
GeoexperimentsResearch - An open-source implementation of the geo experiment analysis methodology developed at Google. Disclaimer: This is not an official Google product.
causallift - CausalLift: Python package for causality-based Uplift Modeling in real-world business
tensor-house - A collection of reference Jupyter notebooks and demo AI/ML applications for enterprise use cases: marketing, pricing, supply chain, smart manufacturing, and more.
Robyn - Robyn is a Super Fast Async Python Web Framework with a Rust runtime.