Skip to content

nurkholiqaganihafid/E_Commerce_Marketplace_Analysis

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

9 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

๐Ÿ“ŒE_Commerce_Marketplace_Analysis

  • Brazilian E-Commerce Marketplace is a real commercial dataset containing information about 100 thousand orders from 2016 to 2018 in various markets in Brazil. Analysis that can be performed includes order status, price, payment, delivery, customer location, product attributes, and customer reviews.
  • The goal of this analysis is to provide valuable insights to e-commerce businesses on optimizing sales strategies, increasing customer satisfaction, and optimizing product delivery.
  • This database consists of nine tables, including customers_dataset, geolocation_dataset, order_items_dataset, order_payments_dataset, order_reviews_dataset, orders_dataset, product_category_name_translation, products_dataset, and sellers_dataset. Dataset source
  • This project uses a data analysis cycle, namely:
    • Question (Business Questions)
    • Data Wrangling
    • Exploratory Data Analysis (EDA)
    • Data Visualization and Explanatory
    • Draw Conclusion and Recommendation
    • Project Result Link: Click here

๐Ÿ“ŒOutline

๐ŸŽฏBusiness Questions

๐Ÿ‘† Business questions that will be answered through this data analysis include:

  1. How are the sales performance, number of ordered items, and revenue on E-Commerce in the last few months?
  2. What products sell the most and least?
  3. How is the distribution of review scores on products sold and what is the percentage for each review?
  4. How is the product delivery time distributed?
  5. What are the demographic characteristics of customers in the e-commerce marketplace sector?
  6. Is there a correlation between product price and shipping costs (freight_value)?
  7. How is the distribution of geolocation based on latitude and longitude in Brazil?

๐ŸŽฏResults

๐Ÿ’ปDeploy Streamlit App

๐Ÿ‘† Requirements to deploy dashboard.py on streamlit.

  • Make sure the virtualenv is ready to run for this project.
  • Install all the required libraries using the following command.
pip install streamlit Babel geopandas matplotlib numpy pandas seaborn streamlit
  • Prepare requirements.txt file
pip install pipreqs

This package is used to generate the requirements.txt file which contains a list of all the packages used in this project.

pipreqs

This command will automatically create a requirements.txt file with a list of all packages used in this project.

  • Run the streamlit application
streamlit run dashboard.py

This is the command to run the Streamlit application with the dashboard.py file.

๐Ÿ“ŠStreamlit Dashboard

๐Ÿ‘† Link Streamlit Dashboard: E-Commerce Marketplace Dashboard

E-Commerce Marketplace Dashboard

๐Ÿ“ˆMy Analysis Results

Merging all Data

๐Ÿ‘† Total Data from the Merge Results

Data Frame

Data Visualization

1st Question

๐Ÿ‘† How are the sales performance, number of ordered items, and revenue on E-Commerce in the last few months?

  • By orders & items

Number of Orders per Month (Last 12 Months)

Based on the visualization above, it can be observed that the month with the highest number of orders is November 2017.

Apart from that, there was a significant drastic decline in September 2018.

  • By revenue

Total Revenue per Month (Last 12 Months)

Experiencing a very drastic decline in the number of orders in September 2018, of course, this will have a huge impact on the E-Commerce marketplace for future profits.

To find out the cause of this decline, it is necessary to consider further analysis regarding several factors, such as the existence of competitors, marketing campaigns, and so on.

2nd Question

๐Ÿ‘† What products sell the most and least?

Best and Worst Performing Product by Number of Sales

Based on the plot above, the Bed Bath Table product has the highest sales.

Meanwhile, Security and Services products have the lowest sales.

3rd Question

๐Ÿ‘† How is the distribution of review scores on products sold and what is the percentage for each review?

  • By number of review scores

Distribution of Review Scores

This distribution gives an idea of how often each review score appears on products sold.

Review score 5 is the one that appears the most with 66,264.

This indicates that the product purchased is very good.

  • By percentage review score

Percentage of Products with Review Score

The highest percentage of review scores is a rating of 5, with 56.5%.

4th Question

๐Ÿ‘† How is the product delivery time distributed?

Distribution of Order Delivery Time

There are some deliveries with a longer time, such as 210 days, which has a frequency of 1. However, this very long delivery frequency is very rare.

Most deliveries occur within a time frame of 1 to 50 days.

5th Question

๐Ÿ‘† What are the demographic characteristics of customers in the e-commerce marketplace sector?

  • By customer state

Number of Orders by Customer State

SP is the state that places the most orders.

  • By customer city

Top 10 Cities by Number of Orders

The city of Sao Paulo has the largest number of orders.

  • By payment type

Number of Orders by Payment Type

Based on the output above, the most commonly used payment method is credit_card. There are also transactions with the payment type (debit_card), which are very few in number.

6th Question

๐Ÿ‘† Is there a correlation between product price and shipping costs (freight_value)?

E-Commerce Marketplace Data Correlation

This output shows the correlation coefficient with other variables in the DataFrame correlation_matrix for the E-Commerce marketplace.

7th Question

๐Ÿ‘† How is the distribution of geolocation based on latitude and longitude in Brazil?

Geolocation Distribution

A geolocation plot has been created by depicting country borders using red color and displaying geolocation points.

The image above shows the SP state that occupies the most geolocation point positions.

๐Ÿ“šConclusions and Recommendations

  • Conclusions

    • Based on the results of the analysis carried out on the E-Commerce Marketplace, there are several important findings. First, sales performance, number of ordered items, and revenue on e-commerce in recent months have been very volatile. November 2017 showed excellent performance, with a high number of orders and revenue. However, there was a drastic decline in September 2018, which had the potential to have a negative impact on future profits.

    • Furthermore, the product that sells the most is the Bed Bath Table, while the Security and Services product has the lowest sales. This shows that products in the Bed Bath Table category have high popularity among customers.

    • In terms of review score distribution, a score of 5 dominates the total of 66,264 reviews with a percentage of 56.5%, indicating that the majority of customers gave very good reviews to the products they purchased. This shows a high level of customer satisfaction.

    • The distribution of product delivery times tends to be in the range of 1 to 50 days, with some other cases of delivery taking longer. However, delivery times that take this long are very rare.

    • In terms of customer demographic characteristics, the state of SP has the largest number of orders, one of which is the city of Sao Paulo which has the largest number of orders. The most commonly used payment method is credit_card, while debit_card has a very small number of transactions.

    • There is a fairly strong positive correlation between product prices and shipping costs, indicating that the higher the product price, the higher the shipping costs. In addition, there is a strong positive correlation between product price and payment value, indicating that the higher the product price, the higher the value of payments made by customers.

    • Meanwhile, the geolocation distribution shows the SP state that occupies the most geolocation point positions, and the state with the least distribution is geolocation_state AP. However, it is important to note that to understand geolocation distribution patterns in more depth, further analysis is required.

  • Recommendations

Based on these findings, several recommendations can be made. First, further analysis needs to be carried out to find out the causes of the decline in performance in September 2018. Factors such as the existence of competitors and marketing campaigns need to be considered. In addition, marketing efforts need to be increased for products with low sales, such as Security and Services.

In terms of product delivery, care needs to be taken to ensure that delivery is carried out on time in order to meet customer expectations. In addition, it is worth considering offering alternative payment methods that are more attractive to customers, besides credit_card.

Back to top

Releases

No releases published

Packages

No packages published