Data Science

Pseudo Facebook Data Analysis

Performed Exploratory Data Analysis through Python and visualization concepts to give valuable insights into data.

Index Of Contents
  ∘ Introduction
  ∘ Data Preparation and Cleaning
  ∘ Exploratory Analysis and Visualization
  ∘ Inference and conclusion
  ∘ Future Work
  ∘ References

Introduction

Data Analysis isn’t one finite set of arrangement of functions with a start and an end. It’s our perspective toward the world around us. It is the collection of data that involves the interpretation of data gathered through the use of analytical and logical reasoning to determine patterns, relationships or trends.

There are 4 kinds of data analysis :
1. Descriptive Analysis
2. Exploratory Analysis
3. Predictive Analysis
4. Inferential Analysis

Source: educba.com

Facebook has a huge amount of data that is available for you to explore. I took the pseudo data from the kaggle for exploratory data analysis.

Data Preparation and Cleaning

1. Gather the data required.
2. Reading the CSV files with some more analytical and statistical analysis like head(), shape(), describe() etc.
3. Cleaning the data i.e. removing the extra data and outliers, filling the missing values etc.
4. Creating the final dataframe for the future analysis

Exploratory Analysis and Visualization

Analysing data through visuals and understand the pattern and trends between features using matplotlib and seaborn library for data visualization.

→ First looking upon the Age analysis, Exploring which age group uses more facebook comparatively and we clearly see teenagers are more using facebook comparatively to others.

→ This graph shows the like analysis which group of people give more likes male or female.

→ Correlational Data Analysis through heatmap

→ To do numerical data analysis, to note the highest and lowest number of likes per day through max and min function and then finding the output

→ Analysing the facebook app users and non-app users with the help of number of mobile likes and likes through the webapp.

→ To analyse the friend count in reference to age groups. Which age-group has more friends.

→ Analysing who has more friends male or female and differentiating through age-groups.

Inference and conclusion

1. The countplot helps in age analysis and showed relation between facebook users according to the age.
2. Scatter plot gives a more enhanced way of differentiating between males and females.
3. Maximum number of likes recorded was 25111.
4. Their are equal amount of facebook web app user and mobile app user.

Future Work

Applying Machine learning and deep learning techniques to future predict the growth in coming years and developing an app of this model.

References

1. https://www.kaggle.com/sheenabatra/facebook-data
2. https://www.kaggle.com/tanyildizderya/facebook-eda
3. https://jovian.ml/learn/data-analysis-with-python-zero-to-pandas
4. https://en.wikipedia.org/wiki/Data_analysis#Exploratory_data_analysis

And so we are at the end of our facebook data analysis. Hope you like this article interesting to get the full code checkout my kaggle link.

Facebook Data Analysis

You can find me on linkedIn :

Anirudhi Thanvi