mehdi-rh / mehdi_rahal

Data Analyst Portfolio

Mehdi Rahal

Portfolio

Project 1: Communicate Data Findings

Investigation Overview

This data set includes information about individual rides made in a bike-sharing system covering the greater San Francisco Bay area. The objectives of this investigation are :

• Explore when are most trips taken in term of time of day, day of week or month.
• Explore trip duration regarding the same last time features, and see if it is influenced by if a user is a member or a casual customer.

Dataset

There are 4702822 trip on the dataset with 13 features ( trip_duration, started_at, ended_at, start_station_id, start_station_name, start_lat, start_lng, start_lng, end_station_id, end_station_name, end_lat, end_lng, bike_ride_id, member_casual ) most variables are object, the trip duration is an 'int' those related to start and end position are 'float', the start time and the end time are 'datetime', and member_casual wich is a categorical variable.

The data can be found here (https://s3.amazonaws.com/baywheels-data/index.html), there is a csv for each month, the data gathered was from 06/2019 to 05/2020
so the the data was concatenated and cleaned by making correct column names, droping erronious data (some observations have a negative trip duration), set some feature datatypes and finally make a columns for day of time, day of week and month of year.

Summary of Findings

In the exploration, the first notification is that the trip duration followed a normal distribution only if we make an exponential transformation, that's why every visualisation for trip duration was done with an exponential transformation. After that a distribution exploration was done for day of time, time on the week and month on the day.

In a second phase : the relationship between variables has been explored, to explore:

Periods with the highest number of trips wich are :

• In term of day of Time : from 08:00 to 10:00 and from 16:00 to 19:00.
• In term of dayof Week : Tuesday, Wednesday ,Thursday and Friday, Monday is slightly lower and the weekend have a low total number of trip.
• In term of Months : February 2020 is the with the highest total number.

The average trip duration : The average is about 10 minutes, some period of time the average is a little high.

• An interesting observation is the negatie relationship between trip duration average and the total number of trips for each period of time.

The influence of the customer type : Casual customers are more susceptible to have a long trip, and have a low total number of trips comparing to member customer.

• This confirms the negative relationship between the trip duration average and the relative total number of trips.
  
  Tools: Python, Pandas, Matplotlib, Seaborn, Jupyter Notebook

Project 2: web scrape amazon product list

In this project I will scrape list of products from many pages with their details: description, price, rating, rating_count

In this project I've used selenium to scrape the rating because it's a hidden field until we click on it.

Tools: Python, Pandas, Selenium, VsCode

Project 3: Gather, Wrangle and Analyse Data

This project is an investigation done on a dataset based on data gathered from the twitter API from WeRateDog page, this is a page that rates people’s dog with a humorous comment about the dog, the investigation is about finding factors influencing the rating and feedbacks quality. In this dataset, we have a lot of information, the best information for our investigation was the rating and the favorite count as feedbacks, and dog breed and dog stage as factors, those two factors are the most influencing feedbacks.

Tools: Python, Pandas, Matplotlib, Jupyter Notebook, Json, Tweepy

Project 4: Analyse A/B test results

A/B tests are very commonly performed by data analysts and data scientists. It is important that you get some practice working with the difficulties of these In this notebook, an A/B test is runned by an e-commerce website, the goal is to know if the company should implement the new page , keep the old page or run the expriment longer to make their decision

Datasets

We have two datasets for this project: ab_data and countries ab_data: comports 294478 rows with 5 features ( user_id, timestamp, group, landing_page, and converted) countries: user_id, country

In this notebook the study is done on three parts:

Part I: for statistical exploration
Part II: A/B Test
Part III: A regression approach

Summary of findings

pval>0.05, zscore<1.64 nb: 1.64 is the critical value for a=0.05 conclusion: we fail to reject the nul hypothesis, so the company should keep should keep the old page.

Tools: Python, Pandas, Matplotlib, Jupyter Notebook

Project 5: Investigate-a-DataSet

In this project a dataset named 'Medical Appointment No Shows' is investigated.
This dataset contains information from 100k medical appointments in Brazil in the 2015 and 2016, the main aspect in this dataset is if a patient show up for its appointment or not.
The main question in this investigation is : What factors are important for us to know in order to predict if a patient will show up for their scheduled appointment

Conclusion

According to my investigation there is no way to predict efficiently if a patient will show up or not.
To answer to the main question wich asks to determine the most important feature to predict if a patient will show up or not: Except the PatientId and the AppointmentID, each feature can lightly help to do the prediction (but with a very low accuracy)

Limitation

The dataset is limited in term of quantity of data, with data of more years the investigation can be more efficient

Tools: Python, Pandas, Jupyter Notebook

Project 6: Sales insights analysis in PowerBI

Objectif: build an interactif dashboard that show the company data in a very clear format to facilitate getting useful insight for better decision making

This project is based on a computer hardware business which is facing challenges in dynamically changing market. Sales director decides to invest in data analysis project and he would like to build power BI dashboard that can give him real time sales insights.

Once sales directory of atliQ hardware has decided to invest in data analysis project he will do a meeting with IT director, data analytics team to come up with a plan. They will use AIMS grid to define purpose and success criteria of this project.

After that, we will look at mysql database that is owned by falcons team. This database has all sales transactions, customers, products and markets information. We will analyse this database and than hook it up with power BI. In power BI we will perform ETL and data cleaning operations to make it ready so that we can build our dashboard.

Then we will plug mysql database with power BI. In power BI we will do data cleaning and ETL (Extract, transform , load). This process is also known as data munging or data wrangling. We will do currency normalization, handling invalid values etc.

Tools: MySql, Sql, PowerBI