Context on the issue
As a Ukrainian from a very young age I was told narratives, to be taken as truth, about people that didn’t look like me, that weren’t white and “European”, didn’t speak my language, or I just had to assume they didn’t. Such narratives told here and there got reinforced over time during school years, silencing the histories of marginalized groups such as ethnic Roma, Hungarian, Jewish citizens. The way history was presented to me was built up on prior textbooks, driven by sociopolitical forces, which in turn perpetuated racist views and othering, contributing to larger issues of segregation, economic inequalities that persist until today.
Coming to TreeHacks I wanted to challenge the views we have formed through our educational curriculum in history, geography, and related fields. History, as it is, is not a mere collection of facts, and is open to interpretation by historians, though it is often taken to be objective, unchallenged. The problem is, most high school textbooks are backed up by other textbooks from the 17th-19th century formed as a series of the same narratives, yet missing pieces of evidence stored, or hidden all over the country at the time. For instance, there is no mention of the majority of Roma population vanishing in concentration camps of 1942, however there exists environmental racism in cities where they reside, with Roma having the least access to healthcare and education, the highest unemployment rates. While bodily narratives exist, they are by no means supported with data and evidence, and can not be justified.
What it does
With observational data now available through history museums, digital archives, private-public collections, urban and environmental data, and accessible for all, the solution I present is called - History Now: Public, Applied, Analyzed, an open source tool for analyzing historical observational data, with users being historians, and researchers in the field including students and professors.
History Now equips researchers with statistical tools for causal inference coded in the backend of the website using R programming software, to establish causality of societal changes, to note trends over time, and critically analyze data proposing changes to improve historical curriculum. The methods used within the tool cover implementation of Rubin Causal Model. The tool includes (1) analyzing observational data as randomized experiments through statistical matching using genetic algorithm (or genetic matching); this allows researchers to compare how the implementation of politics, for instance, affected general population and marginalized groups, (2) examining sensitivity to hidden bias (endogeneity) - this allows researchers to test the robustness of their models for causal inference, seeing how certain level of bias can alter the strength of inference, (3) forming synthetic controls - to evaluate the effect of an intervention in historical comparative case studies, (4) examining regression discontinuity, linear regression and classification, quantile regression.
Given the tools available researchers, and other users can explore and analyze data on the website individually and collaboratively with the team of researchers. They can utilize the certified data set from available archives, input it into the cell, choose variables of interests and causal inference methods. They can test hypotheses, historical and historiographical arguments mostly, to form statistical research papers. A feature of peer reviews for study design will be available, for researchers to refine their hypotheses, test results, and inferences.
The final, peer-reviewed, deliverable can direct researchers to submit governmental petitions for changing the historical curriculum materials used at schools. The recent historical data analyzed can effectively be used to drive policy proposals about allocating resources, improving access to basic needs and rights like healthcare, housing, and education for marginalized minority groups in Ukraine.
How we built it
I made the code for all statistical methods using R Studio; the prototype of the project's website is being created in Figma.
Challenges we ran into
Absence of Web Development specialist, thus I coded some backend for statistical tools and prototyped it in Figma.
Accomplishments that we're proud of
Analyzing the problem and picking the most needed statistical methods to the context of historical scholarship.
What we learned
Context is crucial, and you do have to put yourself into historian's/researcher's shoes to effectively analyze the problem, and the suitability of elements of your solution.
What's next for History Now: Public, Applied, Analyzed
The best is yet to come. Finding a team, doing more historical and statistical research on the topic. A founder (Andriy, https://www.linkedin.com/in/andriy-kashyrskyy/) is actively seeking summer opportunities to get experience in Data Analytics, which will allow him to use some of the knowledge for a side project like this.