Ready Set Research -> 2023 - 2024
River's Research Notebook #4
Replies: 15 comments
-
Week of September 28th Progress ReportI have been using streamlit.io over the past two weeks to build an interactive data dashboard that uses current CDC-verified data on the opioid crisis. Thus far I have been acclimating to using Python as my primary programming language, instead of R, as it integrates perfectly with the data web application I am developing. I have displayed metric data that uses variables and calculations from the data frame I have provided, along with experimenting with Streamlit's graphing and data modeling capabilities. I should most definitely have a functioning prototype in time for next week, although my efforts have currently been diverted to searching for reliable scientific articles and other data frames I can use for my study. I received feedback from OBC about how I can more effectively segment my work, starting by creating figures and seeing what they can tell me about potential patterns in the data. I will start this way, and sure enough, I will find some very interesting trends to explore throughout my research. |
Beta Was this translation helpful? Give feedback.
-
Week of October 6th Progress ReportI have made significant headway with the development of my prototype, as I now have a functioning web app with multiple data visualizations. Within Streamlit.io, I have been able to categorize drug overdoses into their distinct substance groups, allowing for the acute identification of which substances are contributing the most to the current epidemic. Unsurprisingly, Fentanyl and other synthetic opioids excluding Methadone, contributed to 65% of total overdoses in 2022, with death totals increasing 1154.18% since 2015. Using the Plotly, Altair, Streamlit, and Pandas libraries, I effectively plotted these findings through pie charts, comparing the most recent data entries to the oldest. Additionally, these libraries allow for geo-mapping, in which I charted each state's data across a map of the United States. This was incredibly useful, as it allows for an easy visual comparison of each state and its respective contribution to overdose rates across the country. I created two of these maps, one for total overdose deaths, and the other accounting for overdose deaths compared to total population, providing a more accurate understanding of overdose rates per state. Looking forward, I plan to narrow my research to the specific states with the highest overdose rates per population, analyzing the potential reasons why rates are so much higher when compared to states with the lowest overdose rates per population. These factors could include income, poverty levels, access to mental health/rehabilitation resources, stigmatization, criminalization, etc. This will be the foundation of my research and data collection moving forward. |
Beta Was this translation helpful? Give feedback.
-
List of Scientific Research Articles and Other ReferencesBelow is a list of the resources I used for my senior project proposal last semester. I will continue to update this list as I find more research articles that provide scientific evidence for my findings. I will refer to each of these throughout my project as necessary, providing links to each. ReferencesFDA Approves First Over Counter Naloxone Nasal Spray Fentanyl Is Cheaper to Produce than Heroin New Findings on Biological Factors Predicting Addiction Relapse Vulnerability Good Samaritan laws and overdose mortality in the United States in the fentanyl era Adverse childhood experiences predict opioid relapse during treatment among rural adults DEA Reports Widespread Threat of Fentanyl Mixed with Xylazine |
Beta Was this translation helpful? Give feedback.
-
Week of October 12th Progress ReportThis week, as my prototype is functional and serves as a good reference for the direction of my project, I have been researching scientific literature and national events that I can incorporate into my project over the coming weeks. I remembered an article I discovered last year, published in November 2021, that measured the overdose rates from over 3000 counties in the United States that had enacted good samaritan laws. These laws eliminate the fear of punishment or criminalization when it comes to saving lives during potentially unlawful activities, such as illicit drug use. These researchers discovered that overdoses regarding heroin and synthetic opioids approximately decreased by 10% across the board. Theoretically, if the entirety of the United States adopted these laws, that could result in 10,000 lives saved annually. In the next stage of my project, I plan to support these findings through data analysis and make a data-driven argument for policies that favor the decriminalization of illicit drug possession and use. Additionally, I want to continue looking into specific counties where other contributing factors are present, measuring the impact of variables such as poverty, religion, politics, geography, background, etc. I believe this will set a solid foundation to guide the rest of my dashboard's data elements. I will wait for feed feedback on my prototype before moving forward. Goals for the following week:
|
Beta Was this translation helpful? Give feedback.
-
Adding additional research resourcesPercentage of Crime Drug Offenses Per State Both support the findings in my data. |
Beta Was this translation helpful? Give feedback.
-
Week of October 18th Progress ReportIn preparation for my pitch in the following weeks, I have begun developing my slides for my presentation. Thus far, I believe I am making significant progress in the visual aspects of my slides, attempting to be both informative and emotionally engaging as my subject matter is to be taken seriously. My focus is to bring awareness to the Fentanyl crisis that the United States is currently underreporting. In the past several months, news headlines often gravitate towards mass tragedy and loss of life. While the opioid epidemic costs hundreds of thousands of lives annually, you rarely see a headline desperately calling for action. This has become one of my project goals, bringing this societal issue into the spotlight of mainstream media until significant adjustments have been made in the interests of public health. Additionally, I added a few more components to my prototype before submission. I was able to identify another data set containing information on every new rehab facility in the United States since the 1980s. This provides crucial information regarding how different states responded to their overdose rates, highlighting proactive states and those that were not. Along with this, I researched each state's criminalization of substance abuse, comparing them on a spectrum from harsh to forgiving. Interestingly, in agreement with my original hypothesis, I discovered that West Virginia had some of the harshest penalties for drug possession in the nation while also maintaining the highest national fatal overdose rate by population. It wasn't surprising that one of the states struggling the most in the Fentanyl era had some of the most punitive legislation when it came to illicit substance abuse. As if that wasn't enough, using my data analysis, I discovered that in the past twenty years, only two additional rehab facilities were constructed, bringing West Virginia's total rehab locations to a total of nine. It is for these reasons that my senior project strives to illustrate the factors contributing to unnecessary death and the potential solutions that exist within the data. I hope to articulate this further in my project pitch! |
Beta Was this translation helpful? Give feedback.
-
Scientific Literature on the Economic Costs of Opioid DependencyThe economic burden of opioid use disorder and fatal opioid overdose in the United States, 2017 Incredibly useful article with a complete breakdown of the estimated economic costs that the Opioid Epidemic inflicted upon the United States in 2017. This is crucial evidence for my argument, as there are significant financial benefits gained from addressing this evolving public health crisis. |
Beta Was this translation helpful? Give feedback.
-
Week of October 26th Progress ReportFollowing the completion of my prototype, I have been mapping out the future timeline for my senior project. I believe I have an incredibly strong foundation for future research, including measuring and analyzing the progressive economic deficits this public health issue has accrued over the past several decades. If I can't effect positive social change through people's hearts and emotions, I can certainly try through numbers and finances. The next step in my research will be acquiring new government-verified data bases that collect information regarding financial costs, especially in states where overdoses are more common and less treated. In terms of my project's technical aspect, I want to create a drop-down menu that specifies the state my users would want to look at and upon selection, provides state-specific public health information with statistics comparing and contrasting to national averages. This will take a significant amount of time and planning, so it might be useful for me to use tools like Figma to create a wireframe of what I intend to accomplish. At this time in the semester, it would also be a good idea to learn Streamlit's plug-in capabilities and find an appropriate web template to host my dashboard. Goals for next week:
|
Beta Was this translation helpful? Give feedback.
-
Week of November 13th Progress ReportSince my last update, I presented my prototype to the class and received crucial feedback regarding the future of my senior project. My presentation went as well as I had hoped, albeit perhaps less articulate than I would have preferred due to the time restrictions. I believe I succinctly displayed the working demo of my senior project along with describing the problem, solution, and goals of my project. With the presentation completed, I can dedicate the majority of my focus to structuring the rest of my senior project. Following the research conducted over the past few weeks, I established the feasibility of generating state-specific data with user input. The tentative plan is to create a complex algorithm that selects and isolates specific data based on the user request and subsequently generates graphs and information specific to the selected data. This will be incredibly useful when it comes to users wanting to understand the severity of this public health issue within their state and how they may be able to help. This tool can be used as a helpful comparison to national averages, highlighting worrisome states and others that appear to be actively handling the issue. For many, this can help educate people about the effects of criminalization and stigmatization within their home state, as the data analysis conducted by me will provide a solid foundation for policy changes proven to either help or hinder the war against substance abuse and addiction. Additionally, Streamlit has an Embed App feature, which would allow me to create a website with static web builders like Hugo or 11ty and place my data visualizations directly within the website. This allows for greater freedom when designing the user experience and overall web presence. |
Beta Was this translation helpful? Give feedback.
-
Week of November 27th Progress ReportBefore Thanksgiving break, Professor OBC and I discussed the next direction for my project as we head into the next semester. We agreed that my artifact/prototype firmly established my project's feasibility while proving its cultural importance today. The next stage in my senior project includes creating an educational web presence for my analysis, with the target demographic being younger generations. We coined the idea that developing a website that brought immediate attention to the notion of "My taxes aren't going to support your addiction," or something to that effect, would begin the necessary conversation regarding national stigmatization and criminalization of substance abuse disorders. Frequently, many United States citizens assume that their tax dollars are better spent on incarceration than rehabilitation when financial analytics suggest otherwise. According to the U.S. Department of Justice, in 2020 the United States held a total of 1,215,800 people imprisoned while spending about $80.7 billion annually. Breaking those numbers down, that comes out to an estimated $66,000 per prisoner. For comparison, using data recorded by the National Institute on Drug Abuse, the most expensive available treatments available for opioid use disorder, the most popular being naltrexone medication provided in an OTP, including drug and drug administration along with related services, costs a total of... $14,112.00 per year. Now, I am aware of the counterargument that not all incarcerated individuals are imprisoned for drug-related crimes. This is true, however, using studies conducted by drugabusestatistics.org, 46% of federal inmates are serving time for drug-related crimes. With this information, by deducting 54% from the total cost of incarceration per prisoner, the US government spends approximately $35,000 per individual incarcerated for drug-related offenses, or about 2.5 times the cost of the most costly rehabilitation practices. It is through these forms of analyses and comparison that the American taxpayer must begin to question their proclivity to funding punishment over rehabilitation, especially when the costs are several times as large. This argument is the foundation of my planned web development, and the focal point of further analysis to come. I want to provide these clear comparisons to educate young Americans and equip them with the basic economic information regarding addiction in the United States, before they align themselves with political ideals that may be in the disinterest of their own bank accounts. Resources: |
Beta Was this translation helpful? Give feedback.
-
Week of February 5th Progress ReportOver the past two weeks, my primary focus has been on loading and processing new crime and drug-related arrest data obtained from CSV files from government-verified sources such as the FBI and USCC Federal Sentencing Statistics. Through the further development of my Streamlit app, users can now interact with the data, with features such as text inputs enabling the selection of a specific state for personalized data exploration. The data underwent thorough cleaning and processing, addressing issues like column naming conventions and numeric value conversions that interrupted my project's overall functionality. Despite some initial challenges in data processing and user interaction, the app successfully visualizes insights derived from crime and drug-related arrest data through user input. As part of the analysis, I continued to use Plotly Express to create informative pie charts that compared drug violations across different age groups and years. Beyond data visualization, my efforts were directed toward exploring methods for translating state names into their respective abbreviations, adding a layer of versatility to the user input process. This way, the project won't break should a user not capitalize their input or other syntactical differences. Next Steps:
|
Beta Was this translation helpful? Give feedback.
-
Week of February 9th Progress ReportThis past Thursday, OBC and I discussed the project's future direction and catering my dashboard to a specific audience. This following week, I plan on optimizing my visuals so they more effectively display the trends of my data. The crime data analysis at the end of my dashboard will require the most work as it is relatively new and describes several aspects of state-level drug offense prison admissions. Additionally, OBC and I agreed that the information presented on my dashboard may be most impactful with the intent of educating high-school students. As high schoolers have not formed concrete opinions about global issues, like most politicians, my project may be more influential for this demographic. With this in mind, the presentation of my information must be readable and easy to understand for younger audiences. I will avoid using language or terminology that may be considered esoteric for high-school reading levels or individuals with limited backgrounds in statistics and data analysis. I attempted to begin working on the web platform for my dashboard but to no avail, as Netlify struggled to connect to my selected static web-build repository in GitHub. This will be a primary objective for this weekend, as it is essential that I can begin organizing the overall presentation of my findings in an accessible fashion. Finally, I concluded the most recent draft of the Methods of Approach section of my senior thesis. I accurately broke down the research process for my analysis, explaining how my data was collected, how my mathematical formulas were substantiated, and why I selected the visualization methods I used. I plan on proofreading this chapter over the next week, improving overall readability and omitting redundancies in my writing. |
Beta Was this translation helpful? Give feedback.
-
Week of February 19th Progress ReportDuring my ongoing exploration, I began my analytical experiments to measure the relationship between drug violations, prison admissions, and overdose deaths. I began by compiling a large dataset that incorporated data from four separate collections, isolating the year 2022 to provide consistency throughout. I started by conducting a linear regression analysis to identify the statistically significant correlation between drug violations, quantified as 'd.per100', and overdose deaths, labeled 'Data. Value'. Several data points revealed that regions with heightened drug violations per 100,000 people correlated with an increased percentage of admissions to prison for drug-related crime. This linear correlation was substantiated through scatter plots and regression lines that visualized the observed relationship. Additionally, the findings proved to be statistically significant, with P values substantially below .05. The next stage of my experiments will run regression analyses between overdose deaths and their relationship to criminalization rates per state. My analysis has already revealed a potential connection between states with high rates of drug violations and being harsher on drug-related cases in general. Several challenges occurred while attempting to identify these correlations, frequently concerning missing data while ensuring the consistency of the datasets. In the coming weeks, I plan on highlighting specific states, factoring in potential outliers, and weaving together a cohesive and scientifically supported narrative behind my data. |
Beta Was this translation helpful? Give feedback.
-
Week of February 28th Progress ReportDuring our meeting last Thursday, Professor OBC and I discussed exploring the potential relationship between political ideologies and state-level legislation, particularly as it relates to the impact on drug-related crime rates and subsequent incarceration. Our discussion revolved around the observation that varying political affiliations in states have the potential to influence legislative decisions, subsequently shaping the legal landscape and enforcement strategies. I believe that such differences could contribute to fluctuations in crime rates and incarceration trends, which may in turn affect overdose rates and/or access to substance abuse resources. Following our discussion, I analyzed factors such as drug admissions, arrests, and prison populations. I discovered significant correlations between Republican-dominant states and higher rates of incarceration, drug violations, and drug admission percentages when compared to states where the democratic party is dominant. I have shared these results in the experimental results section of my thesis, furthering the debate on whether or not criminalizing drug offenses effectively tackles this ever-evolving public health issue. Despite republican states generally having harsher laws regarding drug-related crime, overdose rates were almost identical between liberal and conservative regions. If criminalizing drug offenses doesn't effectively reduce overdose rates, why incur enormous economic costs? Questions like these will be answered in the final chapter of my thesis, centering around future work and discussions. Next Steps:
|
Beta Was this translation helpful? Give feedback.
-
Week of March 20th Progress ReportDuring my meeting with Professor Luman regarding my senior project, we discussed the significance of incorporating caveats throughout my project and thesis to ensure that data and graphs aren't misconstrued or misused by potential users. They emphasized the need to prevent the misinterpretation of trends and factors unrelated to the crisis, as visualization can be used to suggest trends with factual certainty that spawn from misinformation. Professor Luman proposed the inclusion of a slider feature, enabling users to swiftly compare drug violation rates with a state's political majority, allowing for a deeper understanding of potential correlations between political policies and resulting violation rates. Furthermore, they suggested integrating elements highlighting the theme of time, aligning with the project's title, "Every Five Minutes," to underscore its temporal significance. I incorporated all of this feedback into both my thesis and my project. My project is now equipped with a timer that counts down as users progress through my site, reminding users that the time to act is now while preventable lives are being lost. As the timer strikes 0, a message reveals that one life has been lost since visiting the page, functioning as a visceral call to action. I updated my graphs with slider images so users can click and drag to view one graph layered over another. This allows for further transparency throughout my analytical findings. Each graph is now equipped with a minor description that details potential warnings about how data can be misinterpreted, preventing misunderstandings or potential misuse of my visualizations. Next Steps:
|
Beta Was this translation helpful? Give feedback.
-
My Pitch
Action Items
Notes
Beta Was this translation helpful? Give feedback.
All reactions