PredictBot is a personal project that Identifies arbitrage opportunities in the Political prediction market PredictIt. Using negative risk, it identifies opportunities to hedge against all possible market outcomes, and thus guarentee a profit. FullMarketAccess is a Java class which parses the PredictIt Market API for price information, loading it into internal Market objects which are composed of Contract objects. It then runs a series of simple tests to determine if a Market can be arbitraged. For more on how it works, see the github repo README file. Currently, the most practical way to find arbitrage on Predictit is to examine the sum of “no” prices in large contract markets. In a given market in which one outcome must happen, the sum of all “yes” should add to 100%, implying that sum of all “no” should be (n-1) (100%) where n is the number of contracts in the market. By calculating this value for all Markets, PredictBot identifies when the true sum of “no” prices are less than (n-1) * (100), and returns the market to the user if it is. As an example, If (n-1)100 = 400, and the true market value is 3.96, this presents an opportunity to buy $4 for the price of $3.96.

The Tiny Search Engine Crawls and Indexes webpages, and handles user Queries for relevant information. It was designed and implemented in C with a special emphasis on modularity. The crawler crawls a website and retrieves webpages starting with a specified URL. It parses the initial webpage, extracts any embedded URLs and retrieves those pages, and crawls the pages found at those URLs, but limits itself to maxDepth hops from the seed URL and to URLs that are ‘internal’ to the designated CS50 server. When the crawler process is complete, the indexing of the collected documents can begin. The index module is designed to take as input a page directory and index file name in the format ./index pagedirectory indexfilename This module then outputs the data to indexfilename where each line of the file contains data about word frequency in the form word docID count [docID count].... querier.c is activated by the command ./querier directory indexfile after it is made using make. It is designed to execute from a command line with usage syntax ./querier pageDirectory indexFilename where pageDirectory is the pathname of a directory produced by the Crawler, and *where indexFilename is the pathname of a file produced by the Indexer.

Sudoku is well known logic puzzle where numbers between 1 and 9 are placed on a 9x9 grid of cells. This module is capable of randomly generating and returning a sudoku board with just 1 unique solution to be solved. It is also capable of completely solving a given 9x9 sudoku board if a solution exists. To build this module run command make. For representative test cases, run make test. To make a new random sudoku puzzle, run ./creator [filename] which writes a randomly generated and formatted sudoku puzzle to output file [filename]. To solve a given sudoku puzzle, run ./solver [fileinput] [fileoutput].