Ashley, John Final Presentation

Mini Project/Project Proposal.md

@@ -0,0 +1,33 @@
+# Data Project Proposal - John Park
+
+##### Project Title: Impact of Social Media Exposure for National Park Visits
+
+1. Project Description
+In recent years, social media has played a significant role in shaping travel trends and influencing public interest in outdoor recreation. National parks across the United States have seen fluctuations in visitor numbers, and many studies suggest that social media exposure may be a key driver of these trends. This project aims to explore the relationship between social media activity and national park visitation rates using historical data from the National Park Service (NPS) Visitor Use Statistics Dashboard and external social media engagement metrics.
+- To analyze historical trends in national park visitation and assess fluctuations over time.
+- To investigate the correlation between social media exposure and monthly visitor counts.
+- To identify specific national parks that exhibit significant changes in visitor numbers following viral social media moments.
+- To assess the impact of different types of social media content (e.g., hashtags, geotagged posts, influencer promotion) on visitation trends.
+
+2. Dataset
+- National Park Service Visitor Use Statistics Dashboard: Provides historical monthly visitor counts and visiting purposes (Lodging, Camping, Tent Campers, RV Campers, and Overnight Stays) for various national parks. Such features would include 
+- Social Media Data: Extracted from platforms such as Instagram and Google Trends, including hashtags, mentions, and geotagged posts related to national parks.
+- Additional Contextual Data: Weather conditions, seasonal effects, and promotional campaigns that may influence visitation trends.
+
+3. Methodology
+- Data Collection: Retrieve monthly visitor count data from the NPS Visitor Use Statistics Dashboard and extract relevant social media engagement metrics from available sources with web scraping.
+- Data Preprocessing: Clean and standardize data to ensure consistency, removing anomalies and seasonal trends that may skew results.
+- Correlation Analysis: Apply statistical techniques such as Pearson correlation or time-series analysis to evaluate the relationship between social media activity and visitation trends.
+- Case Study Analysis: Identify national parks that experienced significant surges in visitation due to viral social media moments and assess key factors contributing to the increase.
+- Visualization and Interpretation: Use data visualization techniques to present findings in an accessible format, highlighting patterns and key insights.
+
+4. Affect on Stakeholders
+- National Park Authorities: Can use findings to improve visitor management strategies, allocate resources efficiently, and implement targeted conservation efforts.
+- Local Communities and Businesses: May experience economic benefits due to increased tourism but could also face challenges such as overcrowding and resource strain.
+- Social Media Platforms and Influencers: Findings can help understand their influence on travel trends and encourage responsible tourism promotion.
+
+5. Ethical Implications
+- Privacy Concerns: Ensuring that social media data is collected and analyzed in compliance with privacy regulations and ethical guidelines.
+- Environmental Impact: Addressing potential negative effects of increased visitation, such as habitat degradation, and recommending mitigation strategies.
+- Equity and Accessibility: Ensuring that increased exposure does not lead to the exclusion of certain groups or over-commercialization of public lands.
+- Responsible Tourism Promotion: Encouraging influencers and media platforms to promote sustainable and respectful tourism practices.

Mini Project/Regression Farmworks.md

@@ -0,0 +1,18 @@
+# Project Proposal
+#### Project Title: Optimize Meals and Profits at Dickinson Farmworks
+
+1. Project Description
+Farmworks is a farm-to-table store that prepares meals from fresh ingredients supplied by the Dickinson College Farm. This business serves Dickinson students and faculty and the greater Carlisle area to aim zero-waste emission within the store and strengthen the community. A recent conversation with Jenn Halpin, Director of Dickinson College Farm, suggested that Farmworks is reaching capacity in the number of meals they can produce each day and are therefore running out of food before the business closes. Our goal is to optimize the number of soups and salads the head chef should make, in order to maximize their profits (of each business day 11a to 2p). To pursue this project, we will collect historical data on menu items, amount of food produced and served, cost to produce each meal, timestamp, and potentially weather data to inform the model. Additionally, we will consider the number of workers to scale its influence on the production side.
+We applied linear regression and decision trees for the previous model. I would work more on improving the regression model and coming up with data visualizations with the new features. In addition, possibly time series models. The model will find a balance between the needs of the community and maximize the business profits. We are hopeful that the outputs of the project will inform Farmwork's’ discussion with college leadership to expand budget for another chef and attract more customers by consistently satisfying the demand.
+
+2. Dataset
+We are in the process of collecting data from the Farmworks’ business system (Square) that contains time, date, # of meals served, what they bought (salad, soup, ½ soup, ½ salad or empanadas). We will also receive data on the base estimate cost of each soup and information related to menu rotations and food preparation. We may also choose to include weather data and see how that feature informs our model. For specific features, we plan to include the historical menu data (categorical), number of meals sold (numerical, continuous), special menu items (e.g. empanadas, breakfast, etc.) (categorical), day of the week (categorical), number of servings made (categorical), # of servings sold per hour and estimate cost (numerical), and potentially number of workers/total labor hours (numerical) and weather data (numerical (temp), categorical (cloudy, raining, etc.)).
+
+3. Methodoloy
+- Data Collection: Gathering additional temperature and seasonal data. 
+- Data Preprocessing: Clean and standardize data to ensure consistency, removing anomalies and seasonal trends that may skew results.
+- Regression: Improve the regression model by implementing polynomial regression. 
+- Visualization: Daily sales trends by menu items. Interactive dashboard. 
+
+3. Evaluation
+To evaluate our model’s performance in predicting optimal meal production, we will use regression metrics, including Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared values (R^2). The MAE and RMSE metrics will assist in identifying and adjusting significant prediction gaps. R-squared will determine our model’s effectiveness in capturing key demand features. If it leads to overfitting, we may switch to the adjusted R- squared metric. These metrics will assess the model’s accuracy in predicting the meal quantities, classified by each menu cycle, and effectiveness in meeting the equilibrium point of the meal’s input cost and revenue. After the initial evaluation, we will tune hyperparameters and refine features, especially weather and labor productivity, to improve the model’s performance. The model’s sensitivity to weather and seasonal menu changes will be ensured to maintain a consistent performance.