Angela Detweiler's repositories
NewsSentimentAnalysis
In this exercise I utilized Python libraries - pandas, numpy, matplotlib.pyplot, tweepy, seaborn, datetime, VADER - JSON traversals, and Twitter's API to perform a sentiment analysis on the news mood based on tweets from five different news organizations - BBC, CBS, CNN, Fox News, and New York times.
Improving_Yelp_Ratings_with_ML
Our goal in this group project is to apply NLP and other features from Yelp reviews into a model that outputs a new 5-star-rating, so that there is less discrepancy between reviews and star ratings. In order to make our model more robust, we will also incorporate new user star-ratings based on reviews read (meaning that someone who did not write the review gives a star-rating based on the review text alone) into our model so that it better reflects the review sentiment. We used multiple ML models, including: Naive Bayes, k-NN, K-Means, LSTM, N-Gram, TD-IDF and Linear Regression
CitiBikeTripsNY_Tableau
In this project I used Python to merge multiple .csv files containing Citi Bike Trip History logs. After a quick exploration of the data, I imported the merged file into Tableau and analyzed the use of Citi bikes in New York City from April 2017 through May 2018. I created a workbook, which includes visualizations and a short summary of my findings.
EarthquakeMapping_Leaflet
In this project I utilized leaflet.js to create a map to visualize earthquake locations and magnitudes around the world over a period of 24 hours. The data was pulled from the USGS GeoJSON Feed page.
JavaScriptDOM_UFOTable
In this project I created an HTML web page that contains a table with UFO sightings information that can be dynamically searched through by entering a specific date/time of interest.
PharmaceuticalStudy_Matplotlib
In this study, 250 mice were treated through a variety of drug regimes over the course of 45 days. Their physiological responses were then monitored over the course of that time. I utilized matplotlib, numpy and pandas to analyze the physiological response of mice, under four different treatments (Capomulin, Infubinol, Ketapril, and Placebo).
USCensusData_D3
In this exercise I used data from the U.S. Census Bureau and the CDC Behavioral Risk Factor Surveillance System Prevalence Data to see if there was a correlation between people that work from home (non-commuters) and the frequency of their physical activity. A scatterplot created using D3.js shows the correlation between these two datasets with each circle representing the population from each state in the United States.
ExploringSakilaMySQL
In this exercise I loaded the database 'sakila' into MySQL and ran queries using joins and subqueries when necessary to answer questions about the dataset. See SQL code below for each problem.
WebScrapingMarsInfo
In this project I utilized Jupyter Notebook, BeautifulSoup, Pandas, and Requests/Splinter to scrape information about Mars from several different websites, and utilized MongoDB with Flask templating to create an HTML page with all the information that was scraped.
AdvDataStorageRetrievalHawaiiWeather
In this exercise I used SQLAlchemy, SQLite, and Flask to retrieve data from different datasets, to query, to analyze, and to make the data readily available as browsable api.
artic-ncov2019
ARTIC nanopore protocol for nCoV2019 novel coronavirus
CitySchoolAnalysis_Pandas
In this exercise I utilized Pandas and Jupyter Notebook to analyze district-wide school's standardized test results
ClimateWebVisualizationDashboard
In this project I utilized Python libraries (requests, citipy, pandas, numpy, matplotlib.pyplot, datetime), JSON traversals and the OpenWeatherMap API to collect and analyze weather data from over nine hundred randomly selected cities worldwide (each point in the longitude vs. latitude scatterplot above represents a city analyzed). For all weather parameters analyzed (temperature, wind, humidity, and wind speed) temperature was the most strongly correlated to latitude. The webpage created contains source data as well as visualizations showing the correlation (or lack of correlation) between latitude and four different weather parameters.
craigslist-for-sale-alerts
A Slack/Email bot that helps you find for sale items on Craigslist.
ElectionAndFinancialAnalysis_Python
In this exercise I utilized Python to analyze financial and election datasets.
ExcelPivotTables
For this exercise I reformatted thousands of records in an Excel spreadsheet, utilized conditional formatting, and pivot tables to analyze a subset of Kickstarter campaigns to evaluate which projects are most likely to succeed.
OpenWeatherMapAPI
In this exercise I utilized Python libraries (requests, citipy, pandas, numpy, matplotlib.pyplot, datetime), JSON traversals, and the OpenWeatherMap API to answer a fundamental question: "What's the weather like as we approach the equator?" Weather data were collected for over nine hundred randomly selected cities across the world.
StockMarket_VBA
In this exercise I utilized VBA scripting to analyze stock market data.