In the fourth week, I turned my attention to converting the processed ballot data into a usable HTML format. This required not only accurate text processing but also a focus on user friendly presentation. The iOCR algorithm was again updated to improve the accuracy of text extraction and correction before the data was formatted into HTML. The most challenging part was maintaining the integrity of the data while gauranteeing the final HTML output was clear and navigable. After each update, I tested the results by running new ballot scans through the Brother OCR machine, processing them with iOCR, and then comparing the output against the original ballot data. This process revealed several issues, particularly with the dynamic generation of titles and the handling of larger, more complex ballots. These issues were documented in the spreadsheet, this helped to pinpoint specific areas that needed improvement. By the end of the week, the HTML generation process was functioning.