This thesis addresses the problem of determining the appropriate market
value of a used vehicle. Customers have insuffcient overview of the situation
on the market, so they find it difficult to assess whether the offered price is
reasonable and end up buying a vehicle at too high a price or selling it below
its market value. Therefore, we developed a prototype of a web application for
publishing and browsing vehicle listings, which enables users to sell and buy
vehicles simply and safely. A key part of the proposed solution is a system for
automatic vehicle evaluation, based on collecting data from external sources,
normalising it, and comparing a vehicle with similar vehicles on the market.
The valuation takes into account key vehicle characteristics such as make,
model, year, and mileage. Publicly available data sources and automated
data collection from vehicle sales websites were used to obtain the data. The
system processes the collected data and uses it to compute the estimated
market value of a vehicle. The result of the thesis is a working prototype
of a web application, developed using the technologies React, TypeScript,
Node.js, Express, Python, PostgreSQL, Redis, and Docker. In the analysis,
we used the prototype to show that automatic vehicle valuation is useful, that
users can successfully complete all main tasks, and that the system performs
well under load.
|
