Two years after the first recorded outbreak of Covid-19 its long-term effects are still not completely understood. An unknown proportion of all covid patients go on to develop post-acute covid syndrome and suffer long-term symptoms and health effects long after the initial infection subsides. Project Crowdsourcing Långtidscovid-Sverige sent out in summer of 2021 an open online survey and gathered respondents through crowdsourcing to gather info about people in Sweden with prolonged health effects lasting at least three months after confirmed or suspected Covid-19 infection. In this thesis an explorative analysis of the aforementioned survey is conducted with its initial focus placed onthe progression of symptoms. Descriptive statistics are provided for the survey sample; hierarchical clusteringon principal components is performed; and association rule mining as well as sequence rule mining is used toextract frequently co-occurring symptoms. Women stand for 85.2% of all respondents, possibly indicating a skewed gender distribution in the sample. The average age of a respondent is 50 years old, but ranges between 18 and 80 years of age. The number of reported symptoms tend to diminish over time and symptoms within the 'air passages' category diminish on average quicker than other categories. Hierarchical clustering with Ward’s criterion revealed 4 clusters with an average silhouette coefficient of 0.246. The resulting clusters are not well-separated from each other and have some overlap in their bordering regions, and should therefore be interpreted with caution. Broadly speaking, individuals from cluster 1, 3 and 4 are distinguished primarily by their total number of symptoms reported, meanwhile cluster 2 is characterized by individuals that experience many symptoms early on and fewer symptoms later on. The most prevalent symptom over the entire period is fatigue (90.2%), closely followed by worsening symptomsafter physical activity (87.1%), problems with concentration (82.3%), headaches (79.5%), and brain fog (77.9%). There are several strong associations between various symptoms, especially for symptoms within the same category. Most symptoms have a sequential correlation with themselves and have an increased tendency to occur several times.

This project was originated by K2 atmitec Brno s.r.o. company. The result is data mining module in K2 information system environment. Engineered data module implements association analysis over the data of K2 information system data warehouse. Analyzed data contains information about sales filed in K2 information system. Module is implementing consumer basket analysis.