carVerse

carVerse web app

Inspiration

Making car buying and selling transactions more transparent and equitable was the driving force behind developing carVerse. Knowing an automobile's genuine value has grown more and more challenging as a result of the used car market's significant growth. It can be difficult to precisely evaluate all of these aspects, such as age, condition, and mileage, which all have an impact on a used car's pricing.

What it does

CarVerse aims to deliver price prediction models to the public without the intervention of agents, to help guide the individuals looking to buy or sell cars and to give them a better insight into the automotive sector. The calculator takes in user inputs such as the year, original price, distance covered, fuel type, number of previous owners, and transmission type and uses this information to calculate the car's estimated value. The algorithm uses Random Forest Regressor analysis, which involves analyzing past sales data and identifying patterns and trends in the market to predict the car's value accurately.

How we built it

CarVerse is a machine learning model. So, we used dataset \f cardekho.com from kaggle which consist of 300 rows and 9 columns. All the features are taken for analysis and prediction except the selling price which is our target to find. With the help of feature engineering, The data. corr() gave us an intuition on the correlation between all attributes in the dataset. More correlated features were removed since they led to overfitting of the model. Then, we sliced the data into train and test set. The 'extratressregressor' library allowed us to view feature importances and thereby remove the less important features from the data. We have used the random forest regressor to predict the selling prices since this is a regression problem and that random forest uses multiple decision trees and has shown good results for my model. Then, we finally used the model to predict the test dataset. We use Flask framework for deployment. Then the whole model was packed and imported it into the python file for creating web applications. Hence we dump our model into the pickle file using the given code. Finally the front end was made with HTML and run in annaconda prompt to obtain the required web app.

Challenges we ran into

One of the significant challenges in building a second-hand car price calculator was ensuring the accuracy of the valuations. The accuracy of the calculator depends on the quality of the data used to train the model. Additionally, the machine learning algorithm has to consider many different factors, such as the condition of the car and the local market conditions, which can impact the car's value. Secondly, we had to create environment to run the whole program but we executed them in the existing environment only. So we had to re-do to create a new environment and execute everything.

Accomplishments that we're proud of

creating a user-friendly interface that makes it easy for people to input the necessary information about the car they want to sell or buy. A clear and intuitive design can make the user experience enjoyable and efficient.

What we learned

To build a second-hand car price calculator, we had to learn a lot about the used car market and the factors that affect a car's value. We trained on vast amounts of data related to car prices, makes, years, fuel type, and other relevant information. This training helped us understand the patterns and trends in the used car market and how different factors can impact the price of a car. Secondly, CarVerse pushed us to complete the task in a short period of time by learning and implementing a lot of technologies at the same time.

What's next for carVerse

We believe carVerse has a huge potential for the present and future society in the domain of cars. Over time, we plan to build carVerse to a fully functional app incorporated with live-chat system to interact with the previous owner. Also, we plan to include both buy and sell option where both buyers and sellers can bargain the price and determine their final price as well without the need to pay any extra amount to a third party person/agent. "carVerse" will be the next used car universe.