ribarragi.github.io
Portfolio: Ricardo Ibarra Gil
About me
I am a trained Economist with a strong background in statistics, and a passion for uncovering insights from data. My previous professional experience includes extensive consulting work on energy related policies. During the last few years I have pivoted into data analytics, with special interest in product analytics, consumer behavior, natural language processing, social media analysis, machine learning for problem solving and geospatial analysis. I am currently a Master's student at Carnegie Mellon University, where I have refined and expanded my programming and analytical skill set. At Carnegie Mellon, I have collaborated as a Teaching Assistant for Intermediate Python, Geographic Inforamtion Systems (GIS), and tutored for Introduction to Programming with Python
Projects
Machine Learning
Detection of Monkeypox and simulation of early detection policies using a Convolutional Neural Network and epidemiological models --> Report --> CodeAugust, 2022 - December, 2022
Developed an app to identify monkeypox from skin lesions imagery, as well as an epidemiological model to simulate the app’s impact on the spread of monkeypox in U.S. population
Profiling credit card clients for a bank using dimensionality reduction and clustering--> Report with code
January, 2022 - April, 2022
Analyzed customers' usage of credit cards using Principal Component Analysis and different clustering methods to generate clients' profiles and recommend tailored strategies and products.
Sentiment analysis of social media posts relate to train derailment in East Palestine Ohio: were government, media and people in the same channel? --> Code
January, 2023 - June, 2023
GIS
Access to health services in Mexico City. The role of Medical Offices in Pharmacies (MOPs) --> Story mapJanuary, 2022 - May, 2022
This geo spatial analysis shows that the main logic behind the expansion of MOPs in Mexico City has not been -as often publicized- trying to reach unserved population, but rather capturing a market previously served by other options of healthcare, rather than expanding towards underserved areas. Additionally, state level data provides little evidence about the behavior of MOP doctors towards prescription drugs. But it provides somewhat stronger evidence about some of the factors that drive peoples decision to use their services (lower waiting time, and lower travel times)
Map tornadoes... in the U.S. with clustering analysis --> Map
March, 2022
Analyze historical data of tornadoes in the U.S. from 2000 throu 2008: perform k-means clustering on several measures of impact to identify areas with the most fatalities, injuries, and damages, and determine whether they fall inside or outside of "Tornado Alley", a location of central U.S. where tornadoes are most frequent.
Walkability in two neighborhoods in Pittsburgh --> Map
February, 2022
Assess walkability of two neighborhoods in Pittburgh, PA, based on different criteria such as density, availability of parks and playgrounds, terrain elevation, and street patterns.
Distribution mapping for Electric Vehicle Charging within urban communities
January, 2023 - June, 2023</br> Locate the best sites for Electric Vehicle public charging stations, considering current and potential demand, equitable access, physical feasibility, sociodemographic factors, and walking distance restrictions in urban communities of the state of Pennsylvania, mainly using R, Python and ArcGIS.
R and R Shiny
Analysis of car crashes in Allegheny county --> Report --> CodeDate, 2021 - Date, 2021
Exploratory data analysis of car crahses in Allegheny County, exploring potential factors that increase the risk of high impact events.
Crimes in Chicago, 2022 --> Dashboard --> Code
Date, 2021 - Date, 2021
Dashboard showing crimes committed in Chicago during 2022. The dashboard contains: 1) a bar chart with the number of crimes per month; 2) a tree map with the type of crimes; 3) a heat map witht he crimes by time and day of the week; and 4) a table with the raw data. All 4 elements are reactive to the filters: Location of crime, include/exclude domestic violence crimes. Selected data can be downloaded as a csv file.
U.S. airports (1990 - 2022) --> Dashboard --> Code
Date, 2021 - Date, 2021
Dashboard displaying 1) location of U.S. airports on a map; 2) stacked bar chart with total flyers in selected airports by year and by carrier origin; 3) top 5 carriers operating in selected airports with total no. of passengers. All elements are reactive to the filters that enable the user to select: the airport(s) to be shown in map and data dsiplayed, the years for which the data will be displayed; 3) the state(s). Selected data can be downloaded as a csv file.
Shopping store data --> Dashboard --> Code
Date, 2021 - Date, 2021
Dashboard contains three bins: 1) products sold and customers, shows information about cusomter's income, amount spent per type of grocery and composition of sales; 2) sales channel, displays data about sales channel (online, in store, catalog) over time; 3) data table, shows all selected data as a raw table, which can be downloaded in .csv format. All 3 tabs contain graphs and tables reactive to filters that enable the user select: customers' age range, customers' education level; number of children in customers' households.
Other projects
Interactive covid vaccination maps in the U.S. (plotly) --> Continental U.S. --> Florida --> Georgia --> PennsylvaniaAugust, 2021 - December, 2021
Education
Carnegie Mellon UniversityMSc in Public Policy and Data Analytics
ITAMBA in Economics