submit idea 1

Submission Folder/Idea1DaultonKendall.md

@@ -0,0 +1,13 @@
+## Idea 1 
+##### Daulton Kendall
+##### 1/29/2025
+#
+#
+#
+To break down the fair price valuations for existing housing markets, there needs to be a new level of transparency within the preexisting pricing scheme that has gripped current and potential homeowners. In looking to create clarity in the real estate market where historically the power has been placed within corporations and realtors, this model looks to shift the power dynamic back into the buyers’ and sellers' hands. The model looks to expose the correlation between housing features and price and reveal how they influence the fair price valuation for a house. This can also be used to understand the valuations of existing houses to see if they are undervalued or overvalued.
+
+In creating this model, a variety of locations scraped through Realtor.com’s HTML code helped create a vast database of different houses and their features. Using Variables such as List_Price, beds, baths_consolidated, sqft, and location-based variables, the database can paint a picture of different housing markets as well as convey patterns in the differences between the influence of housing features on price in different markets. The model calculates feature importance using a tree-based model such as a random forest model. Utilizing this modeling approach will allow for a more comprehensible data output to buyers and sellers, highlighting the most important features in a digestible manner that can be more readily used in visual representations. This type of modeling will help buyers and sellers understand what affects home prices the most while mitigating factors like less important variables that could lead to overfitting. This also improves the model's performance on larger scalings as the model looks at strong predictors and can reduce computational cost.
+
+The Real Estate Feature Importance Model aims to guide its stakeholders and for this model, that is the buyers and sellers. The model looks to inform buyers and sellers, creating more transparency within the real estate market and shifting the power out of the hands of solely corporations and real estate agents. Buyers and sellers will now better understand whether houses are fairly valued given their key features. This will help buyers and sellers have more informed financial decision-making around real estate, especially for homeowners trying their best to boost their house value for their given housing market. Moreover, for buyers, they will now be able to negotiate overvalued listings with data-driven insights to support their negotiations. These insights are all driven by data provided through their given market, making each market analysis unique and can provide accurate comparisons for houses within that market, compared with one another.
+
+With all models, there are ethical implications that can taint the model's outputs. Biases amongst housing markets especially can further discriminatory housing disparities and can perpetuate discrimination against lower-income areas, even if unintentional. Models like these can reinforce socioeconomic inequalities and can undervalue the homes of minority communities. What furthers the complexity of these ethical implications within the modeling is the fact that the data is derived from Realtor.com, where these biases may have already been a factor in the pricing models. One way to help mitigate these biases is through strict data governance standards during the collection process.

Submission Folder/Idea2DaultonKendall.md

@@ -0,0 +1,19 @@
+#
+## Idea 2 
+##### Daulton Kendall
+##### 2/6/2025
+#
+#
+#
+Trying to understand the volatile markets of cryptocurrency, this model looks to utilize new sentiments from reputable news sources to compile a sentiment analysis on a specific currency. This model looks to create transparency within the cryptocurrency space and create more informed investors. This model will utilize past data to gauge the accuracy of the predictions and then create new predictions on current market sentiment analysis.
+
+The data will be scraped from news sites like Yahoo Finance and pulled articles on the cryptocurrencies ticker. The dataset will compile headlines, time stamps, and content and review them for their sentiment. It will then be fed historical data on the cryptocurrencies prices which will allow the model to learn if it was accurate in its predictions. The data will be sourced through selenium and will be cleaned before being fed into the model.
+
+Utilizing a FinBert model as well as a random forest model will give insight into the sentiment of news articles using sentiment tags such as positive, neutral, and negative. These models will be judged using accuracy, recall, precision, and F1 score to understand the effectiveness of the model.
+
+The model can benefit traders as well as financial analysts by providing a deeper insight into news sentiment on a broader scale. It looks to give not only current news information but also historical information to see any patterns in news sentiment and its correlation to price. It can aid in trading recommendations and can help investors use sentiment analysis in their risk assessment strategies of the currencies.
+
+In scraping news data, there are ethical and legal implications regarding data ownership, however, the data here is public information without financial benefit, making our use compliant with Yahoo Finances policies. Additionally, editorials and other information in the modeling could lead to errors in output creating misinformation, and have the potential to inform poor financial decision-making. Addressing the bias in the training data will be crucial to maintain transparency in the model's decision-making and help mitigate any risks that come with the model.
+
+In leveraging web scraping and news sentiment models, our sentiment analysis model looks to provide helpful insight into understanding the complexities of the cryptocurrency markets. This model looks to continue to grow and provide more fine tunes and live updates in the future, breaking down the hierarchy of information within the cryptocurrency markets.
+