Hi data scientists 👋👋👋.A coincidence that the day we just finished was Valentine’s Day but I have been receiving a lot of love 💙 from you guys lately! Many of you have reached out and supported the content, just know that it’s appreciated and it will make this page better! 👌.With that, let’s get to today’s post, shall we?? .Of course when we are talking about Pandas, our good friends the founders are back to help us! 🎉.However our founders have been having a little argument lately. Even one was allegedly heard calling another one “hey boomer” ‼️ and the other responded back with “you millennial” 🤦🏻‍♂️ Even though arguments are not nice, this gives us the chance to use Pandas to settle who is in what generation!.pd.cut allows us to categorize a continuous spectrum into bins 👌 here our bins are the generations and the continuous spectrum is the year number 👍.After seeing exactly who’s in what generation, our founders realize that they should apologize to each other. They have promised to treat each other better in the next post so stay tuned 🗣.Correction: founders_df[“Birth”] should be founders_df[“BirthYear”] ‼️‼️

.

👨‍💻#Pandas

Hello all and welcome to the post of the day 🙏 .Today, we are going to introduce Machine Learning and Deep Learning and talk about what makes them different 🤔 .In traditional machine learning, scientists had to define concrete and well defined features for the inputs, those features would then get fed into a neural network that would produce a prediction 👍.In deep learning however, we are leaving it to the network to learn and ultimately decide which features it seems relevant to the learning problem 💡 .This is precisely why deep learning is so powerful, everything end to end is learned by the network. The hard part then becomes designing the perfect network for a given problem 🧠.Super excited to be going through this journey through AI with you guys. Stay tuned for more machine learning posts this coming week 🎉

____

👨‍💻 #Machine_Learning

Hello data science enthusiasts 👋👋.Weekend calls for a machine learning related post, doesn’t it?.Machine learning historically started with the two main types: supervised and unsupervised..Overtime, a new type was invented “reinforcement” learning, and now there is even more types ....So, what are they?.Supervised: training a model with labeled data points, you “supervise” the model by giving it the “right answers”.Unsupervised: you ask the model to tell you what it thinks the data classifications or clustering should be based on the pattern it can find in the data. This is a good approach for when there are no right answers or the right answers are not available..Reinforcement: this type is largely evolving and generally is orchestrated on a series of actions and rewards. The model learns over time what action to take and when to optimize its total rewards..Machine learning is fast moving field and the research in it brings a ton of new ideas every month 👌🧠.We should be covering the different techniques used on this slide in the future posts so stay tuned 🗣🗣

____

👨‍💻 #Machine_Learning

Hello everyone 👋👋.Today we are continuing our journey in machine learning with this 3rd post in the series 👍.Logistic regression is one of the most widely used and popular classification algorithms out there. Due to its diversity, simplicity and robustness, it’s become super popular as a baseline model all along the field 👌.At the heart of logistic regression, is the sigmoid function, a smooth function that takes any value and outputs a value between 0 and 1. This function allows for any input to be “classified” in one of the two binary classes after a threshold is applied 👏.Neat, right? 😍

____

👨‍💻 #Machine_Learning

Hi everyone 👋👋 and welcome to another “Intro to Machine Learning” post 🧠.Supervise Learning is everywhere. In fact, 90% of the problems I have solved so far with ML have been through Supervised Learning.With Supervised Learning, you can answer so many questions and become an expert in ML 👌.The two types of Supervised Learning are crucial to Artificial Intelligence: regression and classification.An example of regression is predictions the price of a home based on its number of bedrooms, number of bathrooms, size and age 🏠 ..An example of a classification problem could be predicting whether a cancer tumor is benign or not 💊 .

____

👨‍💻 #Machine_Learning

Hi everyone 👋👋.Hope you’re having a nice weekend 👌.Today we are continuing our journey with the “intro to machine learning” series.And of course, SVMs 🎉.SVM or support vector machine is a machine learning algorithm whose job is to find a hyperplane that divides the data for the each label 👍.The power of SVMs comes through especially when the data is non linear or not easily distinguishable by eye!.SVMs use non linear kernels to transform the data to another space where the resulting data points are easily divisible by a hyperplane and then they transform everything including the hyperplane back to the initial dimensions.How cool is that with the right kernel, the SVM in the post was able to figure out the circular pattern of our data points? 👏👏.I personally think it’s absolutely amazing!!.Let me know what you think below.

____

👨‍💻 #Machine_Learning

What do you say we learn some machine learning today? 👌🎉.Linear regression is typically one of the first machine learning algorithms we hear about when we start learning about ML in general.Disguised in a simple intuitive algorithm, there exists a series of foundational concepts in ML 🤔.To find a loss function whose minimization gives the problem the optimal result.To visualize that loss function and how to minimize it 👌👌.Did you know that companies such as Zillow and Redfin have used a flavor of Linear Regression to predict prices of homes?!! 😱
____

👨‍💻 #Machine_Learning