
Data Analyst & Scientist with 4 years of experience. Sharing LeetCode solutions, data insights, and the latest in tech. #freelancer #dataconsultant
Joined March 2024
- Tweets2.4K
- Following561
- Followers10.6K
- Likes1.7K
Deeksha retweeted
Google Maps just received massive updates.
AI has taken over, and people can't stop getting creative with it.
10 incredible features:
Deeksha retweeted
#Day11 of how I approached data science as a beginner
➡️ I used to think how we can use different features with different range of values in one formula. I tried fitting a regression model without knowing about scaling techniques and the model was behaving crazy.
➡️ However after a lot of reading in the internet, I came to know about feature scaling
➡️ Since data science is mathematics behind the scene, calculations plays an important role.
➡️ Consider two features - Age and Salary. Age range is 0-100 years and salary can range from 0 to anything. When we use these two features in the calculations, the results will be inflated.
➡️ To avoid such issues, feature scaling is done. the main idea behind feature scaling is to bring all the features in the same range (0-1, -1-1, etc)
➡️ Commonly used Feature scaling techniques - MinMaxScaling and StandardScaling
#Day10 of how I approached data science as a beginner
🔷 Feature Engineering is one of the most important aspect of data engineering.
🔷 I spent a lot of time to understand how we can create features from the dataset. The features that we create should have some relation with the target variable.
🔷 There are two kind of features - Categorical and Continuous. For categorical feature One hot encoding is performed as it converts the categorical features to numerical values.
🔷 For continuous features, different techniques like mean, median, mode, standard deviation, moving averages, etc. is used based on the use case.
One should spend 70% of the time in feature engineering.
Deeksha retweeted
#Day6 of how I approached data science as a beginner
Head First Statistics Continued…
Once I covered these concepts I was pretty confident that I am close to completing the statistics required for data science.
🔷 One thing to understand about data science is that you are not dealing with one individual when you train your model. You have to lookout for patterns across the population.
🔷 Dataset follows its own patterns which is called as distribution in a sophisticated terms
🔷 There are different type of distributions like Discrete Probability Distribution and Continuous Probabilty Distribution
🔷 Binomial, Poisson, Geometric Distributions are example of discrete distribution
🔷 Normal Distribution is a continuous distribution
🔷 It is very important for a data scientist to understand the nuances of these distributions and where to use it
#Day5 of how I approached data science as a beginner
🔷 Once I was comfortable with the basics of statistics concepts like mean, mode, standard deviation, etc. It was time to move to the advance concepts
🔷 Machine learning at the end of the day is about probabilities. A lot of concepts in machine learning is based on probability. All of us have read probability in high school and engineering, its just that we have to brush up those concepts
🔷 The chapter 4 of head first statistics helps us visualise probability. Its not just about solving equations but the feel of how probability is used in predicting different classification use cases is a great feeling
🔷 This chapter also talks about one of the favourite theorem of everyone - “The Bayes theorem”
Stay tuned for more..
Deeksha retweeted
5 steps I followed in my career to stand out -
1. Take ownership
2. Keep up with the new trends in your tool stack
3. Try to expand you skills horizontally -
e.g. -
i. If you are a data analyst, learn data engineering, and data science.
ii. If you are a web dev, become a full stack.
4. Work for a startup for some years as it helps for holistic career development
I worked in a startup for 4.5 years after leaving an MNC
5. Try to train a set of people in your area of expertise. It helps to improve your presentation skills as well as others get benifitted from you.
I have trained more than 500 people through out my career!
Keep hustling!









