Pain Points & Pull Requests

Shownotes

In this episode of Pain Points and Pull Requests, Fatimah and Carla discuss all things habits. They go over their favourite takeaways from books, The Power of Habit by Charles Duhigg and Atomic Habits by James Clea. Join them to hear about some of their learnings, tips and ideas to improve their day-to-day habits.

 

Time Stamps

1:20 What makes up a habit?

2:30 Key takeaways from Atomic Habits and The Power of Habit

17:01 Habits we want to implement in the future

 

Resources

The Power of Habit

https://www.amazon.ca/Power-Habit-What-Life-Business/dp/0385669763/ref=sr_1_1?dchild=1&keywords=the+power+of+habit&qid=1620594383&sr=8-1

Atomic Habits

https://www.amazon.ca/Atomic-Habits-Changes-Remarkable-Results/dp/B07GBKBL1B/ref=sr_1_1?dchild=1&keywords=atomic+habits&qid=1620594421&sr=8-1

Atomic Habits Book Summary:

https://www.samuelthomasdavies.com/book-summaries/self-help/atomic-habits/#:~:text=The%20Four%20Laws%20of%20Behavior,hand%20that%20shapes%20human%20behavior.

In this episode of Pain Points and Pull Requests, Fatimah and Carla try a deep learning approach to predicting views on a youtube video. Deep learning is a subset of machine learning, where using artificial neural networks, systems can learn patterns in data to predict outcomes. Fatimah and Carla discuss how they approached experimenting with deep learning, whilst  trying to improve the performance of their linear regression model. 

 

Timestamps

0:40 Part 1 overview

3:10 What is deep learning? 

6:40 Quantifying model improvements and network topology

 

Github project

https://github.com/CarlaLeal/Youtube-Video-Views-Prediction

Shownotes 

In this episode of Pain Points & Pull Requests, Fatimah and Carla break down the components that make up data ecosystems. Data has become a core strategy component for many companies: large and small. For this episode, Fatimah and Carla explore a framework for understanding how technological ecosystems are designed, in order to support a company’s data needs.

 

Timestamps

1:27 What is a data ecosystem

2:10 Breaking down infrastructure, analytics and applications

6:30 How to approach designing a data ecosystem, a walkthrough

 

Resources

What is a data ecosystem

https://mixpanel.com/topics/what-is-a-data-ecosystem/#:~:text=A%20data%20ecosystem%20is%20a,%2C%20operations%2C%20and%20marketing%20decisions.

 

Understanding big data: the ecosystem

https://dataconomy.com/2014/06/understanding-big-data-ecosystem/

Show Notes

In this episode of Pain Points and Pull Requests Fatimah and Carla discuss their key takeaways from Never Split the Difference, a book on effective negotiation. Learning how to negotiate effectively can bring many benefits to both your work and personal relationships. Recognizing this, Fatimah and Carla decided to breakdown some of the most useful  techniques they learned about in this latest read. 

 

Time Stamps

0:47: What is negotiation

1:38 Why care about negotiation?

3:08: Key takeaways from Never Split the Difference

 

Extra Resources 

Book concept cheat sheet

Never Split the Difference by Chris Voss with Tahl Raz

Shownotes

In this episode of Pain Points and Pull Requests, Fatimah and Carla give an intro to Long Short-Term Memory models and some of their useful applications. In addition to breaking down some concepts, they give an overview of a fun project they completed, generating valentines day messages using PyTorch and LSTMs.

 

Timestamps

2:14 An intro to LSTMs

10:20 Applications of LSTMs

12:10 Generating Valentines Day messages using PyTorch and LSTMs

 

Extra Resources

Why use LSTMs?

https://www.carlaleal.net/post/a-blog

Understanding LSTM networks

https://colah.github.io/posts/2015-08-Understanding-LSTMs/

Text Generation Tutorial with PyTorch

https://www.kdnuggets.com/2020/07/pytorch-lstm-text-generation-tutorial.html

Valentines Day Message Generator

https://github.com/CarlaLeal/valentines-message-generator

Shownotes

In this episode of Pain Points and Pull Requests, Fatimah and Carla walk you through their term projects for an undergraduate data mining course. Carla discusses the work her team did to predict whether a homicide will be solved or not and Fatimah discusses how she and her team analyzed the effects of COVID-19 on homeless shelter occupancy.

Timestamps

0:45 Part 1: Predicting Whether a Crime Is Solved Using ML

3:23 Data Analysis

7:00 Working with Imbalanced Datasets

10:21 Decision Tree and Logistic Regression Predictions

13:22 Part 2: Analyzing the Effect of Covid on Homeless Shelter Occupancy

18:52  Data Analysis and Data Preparation

22:17 Cofactor Analysis

26:41 Associated Rule Mining

 

Extra Resources

Techniques to Handle Imbalanced Datasets

https://towardsdatascience.com/the-5-most-useful-techniques-to-handle-imbalanced-datasets-6cdba096d55a

Source code for project #1: predicting whether a crime will be solved:

https://github.com/CarlaLeal/Predicting-Whether-a-Crime-Will-Be-Solved

Source code for project #2: 

https://github.com/FatimahAreola/TorontoSheltersAnalysis

CBC article: City has far fewer homeless shelter beds than it says it has

https://www.cbc.ca/news/canada/toronto/toronto-shelter-space-1.5808905

University of Calgary study weather effects on homeless shelter occupancy

https://journalhosting.ucalgary.ca/index.php/sppp/article/view/42500/30390

In this episode of Pain Points & Pull Requests, Fatimah and Carla discuss their process and findings from working on another Kaggle competition. This time, they tackle how to build a recommendation system to pair potential donors to a project they could be interested in donating to. 

 

Timestamps

1:12 What is a recommendation system?

2:12 Exploratory data analysis

10:26 Content-based recommendation system

12:29 Evaluating recommendation systems

14:47 Collaborative filtering recommendation system

Extra Resources

Donors Choose Kaggle competition

https://www.kaggle.com/donorschoose/io

An introduction to recommendation systems by Google

https://developers.google.com/machine-learning/recommendation

Donors For Good exploratory data analysis

https://www.kaggle.com/fatimahareola/data-exploration-charity-recommendation-system 

Charity donation recommendation system using scikit-learn and tensorflow

https://github.com/CarlaLeal/Charity-Donation-Recommendation-System

TensorFlow Recommenders

https://blog.tensorflow.org/2020/09/introducing-tensorflow-recommenders.html

DataKind

https://www.datakind.org/

Shownotes

In this episode of Pain Points and Pull Requests, Fatimah and Carla break down a product management interview question Fatimah was given a few months ago: “How do you design a fridge for someone who is visually impaired?”. Join them to learn what  techniques you can apply to solve tricky design questions on the spot.

 

Timestamps

0:40 A real-life product management interview question

4:50 Breaking down the question

16:50 brainstorming solutions

27:03 Critiquing and Prioritizing solutions

Extra Resources

Episode 14: An Introduction to Product Management with Daivik Goel

https://open.spotify.com/episode/6sqciFQekF8i9PNOPxcOw0?si=VYOvJXwfTZqgeKvyVix-UQ

Episode 6: Study Session with the Interns

https://open.spotify.com/episode/5c3n3MALEMSwdSwIOmRqvF?si=IqIhsUDqRnKdH-je4QYs9w

In this episode of Pain Points and Pull Requests, Fatimah and Carla discuss how they sliced and diced Youtube trending page and channel data to generate some interesting insights. They go over what targets they’d use to measure their success in producing content based on what they learned from their exploratory data analysis of these datasets. They also talk about the model they produced and what they did to make it perform better than other models used to predict the same target variable.

Full shownotes:

https://github.com/LumakiLabs/painpoints-pullrequests/blob/master/episode-19.org

----more----

In this episode of Pain Points & Pull Requests, Fatimah and Carla discuss what they've learned about how to practice new skills effectively. If you're also preparing for recruiting season, you might hear about some new techniques for breaking down difficult skills and dividing your time effectively to practice them.

Resources

 

Deep work: https://www.amazon.ca/Deep-Work-Focused-Success-Distracted/dp/1455586692/ref=sr_1_1?dchild=1&keywords=deep+work&qid=1600212572&sr=8-1

Essentialism: https://www.amazon.ca/Essentialism-Disciplined-Pursuit-Greg-McKeown/dp/0753555166/ref=sr_1_1?dchild=1&keywords=essentialism&qid=1600212593&sr=8-1

A Mind for Numbers: https://www.amazon.ca/Mind-Numbers-Science-Flunked-Algebra/dp/039916524X/ref=sr_1_1?crid=CMWR9JPBD45F&dchild=1&keywords=a+mind+for+numbers&qid=1600212612&sprefix=Mind+for+Number%2Caps%2C152&sr=8-1

Rewriting the Code(Finding mentors that are women in tech): https://rewritingthecode.org/

Time Stamps

3:47 Skills we're learning

9:05 Prioritizing our learning

19:34 5/25 rule

20:42 Deliberate practice

23:09 Tips for practicing deliberately

31:12 Book recommendations