Inspiration
As college students passionate about the stock market, my friend Yash and I wanted to start investing—but quickly realized how difficult it was to make informed decisions without solid research. We were overwhelmed by how scattered the information was: news articles, financial metrics, and technical charts were all on different platforms. On top of that, most students (including us) don’t have a lot of money to lose, so accuracy is critical. We knew there had to be a better way. That’s when we came up with the idea to build an all-in-one platform that combines fundamental analysis (news, trends, sentiment) with technical analysis (price history, simulations) to help users—especially students—make smarter, more confident investment decisions.
What it does
The Financial Analysis & News Summarization System is an interactive tool that walks users through the full process of researching a stock. Here’s what it does step-by-step: Asks for a stock ticker from the user.
1.Pulls 50 recent news articles related to that stock using NewsAPI and web scraping tools, and displays short descriptions of each.
2.Allows users to select any number of articles to be summarized using Hugging Face’s BART model, providing AI-generated summaries to help users understand recent sentiment and trends.
3.Moves on to technical analysis, showing the stock’s historical prices and calculating measures of central tendency (mean, median, mode) to give users statistical insight.
4.Generates and visualizes a Monte Carlo simulation with over 30,000 projected price paths, allowing users to explore a range of future outcomes based on historical volatility.
- All in all, the system gives users a 360-degree view of the stock they’re researching—news, data, and predictions in one place.
How we built it:
We developed the project entirely in Python, using a variety of powerful libraries and tools: *pandas, numpy, and matplotlib for data manipulation and visualizations *yfinance and Yahoo Finance API to retrieve historical stock prices and real-time financial data *NewsAPI, requests, and BeautifulSoup for gathering, parsing, and cleaning news articles *newspaper3k (with lxml_html_clean) for clean and readable article extraction *Hugging Face’s BART model from the transformers library to summarize selected articles using a state-of-the-art neural network *Monte Carlo simulations were implemented to simulate future stock prices using probabilistic modeling based on historical volatility
We designed the project with reusability in mind, allowing us to expand the code later and test the prediction accuracy over time.
Challenges we ran into
*One of the biggest challenges was finding reliable and free APIs. Many APIs had limited quotas or required paid subscriptions, so we spent a lot of time comparing features, limitations, and pricing to stay within budget while still providing high-quality data. We also faced time constraints since this was a project we had partially started before, but left unfinished. Picking it back up and expanding the system took longer than expected. Debugging API-specific syntax was tricky—especially when small naming differences (like len vs. length) caused errors that were hard to trace. We had to manage large data sets and model outputs, especially when summarizing multiple articles and generating 30,000+ simulation paths
Accomplishments that we're proud of
We successfully created a tool that can summarize dozens of news articles using advanced NLP and serve them to the user in a way that’s quick and digestible. We implemented Monte Carlo simulations with thousands of price paths, providing realistic projections of a stock’s future behavior. The final product delivers a powerful yet simple experience, helping users research a stock comprehensively without needing to jump between platforms. We made the platform free, scalable, and student-friendly, keeping costs low by optimizing around open-source tools and APIs. Ultimately, we’re proud that we created something useful for ourselves and our peers—a decision-making assistant for investing smarter
What we learned
We deepened our understanding of financial modeling, especially in building and interpreting Monte Carlo simulations for stock price forecasting We learned how to work with large language models (LLMs) and optimize them for summarization tasks using Hugging Face’s tools. We improved our skills in web scraping and API integration, developing clean workflows for merging news and financial data. We also learned how valuable organization and documentation are—especially when dealing with multiple libraries, data pipelines, and changing syntax across APIs. Most importantly, we learned that bringing together fundamental and technical analysis in one place creates a much clearer and more powerful picture of a stock’s potential.
What's next for FINANCIAL ANALYSIS & NEWS SUMMARIZATION SYSTEM
We’re excited to keep building on this project. Our next goals are: Track selected stocks over time to compare our Monte Carlo predictions with actual performance and refine our simulation algorithms. Package the system into a web-based interface or mobile app so more users (especially students) can access it easily. Implement sentiment analysis to detect whether recent news is positive or negative, adding another layer of insight to the summaries. Introduce portfolio recommendations or alerts based on user risk preferences and market trends. Eventually, allow users to save their research sessions, revisit past simulations, and even simulate portfolio growth over time. We built this project to help us invest smarter, and we’re excited to continue evolving it to help even more people do the same Our big goal is to create a UI which will give it a much cleaner and smoother look for our users
Log in or sign up for Devpost to join the conversation.