Visualizing IVF Patient Journeys
            
            
            MAT 259, 2023
            
            Jenni Hutson
            
            
 
            
            
            Concept
            
                For my final project, I was interested in visualizing data which 
                represents very personal decisions or times in people's lives.
                These topics seemed more difficult to reduce down to numbers, so
                I was curious if I could create a visualization which represented
                them well. I ended up finding a dataset which included 160,000
                patients' data who underwent fertility treatment. I was specifically
                interested in patients who had IVF treatment (In vitro fertilization). 
                IVF is a medical procedure whereby an egg is fertilized by sperm in 
                a test tube or elsewhere outside the body, and is used by people
                who are unable to concieve naturally for a variety of reasons. 
                
            
            
 
        
            Query
            
            
            
            Preliminary sketches
            
                IVF can be a long process, and many patients undergo multiple
                rounds before they are able to concieve. I wanted to create
                a sort of timeline which showed each patients journey through
                the various stages of treatment.
    
                The stages could included   
                
 
                    - Number of eggs mixed
                    
 
                    - Number of Embryos created
                    
 
                    - Number of Embryos transferred
                    
 
                    - Early pregnancy outcome: whether Intrauterine Fetal Pulsation Seen, 
                    biochemical pregnancy only (early miscarriage), or miscarriage  
                    
 
                    - Number Weeks pregnant
                    
                    - Pregnancy outcome: live birth or no
                
                
 
                
            
        
            
            Process
                
                The initial dataset had ~130,000 rows of patient data who underwent
                IVF. To select a more mangeable subset, I took a random sample to get
                about 3000 rows. With the random method, there were far more patients
                with male partners than patients with female partners or no partners.
                Since I was curious if there would be any variations in these three groups,
                I changed my method to select about 1000 patients from each group. 
                Within each group, these patients were randomly selected.
                
            
            
 
            
             
            Final result
                
                For the final project, I mapped each stage to a rectangle. If 
                the data for a stage wasn't numerical, I designated steps on the
                y-axis to represent each category for that stage.
                Many patients did not have a child as a result of this round of IVF.
                Since their lines went to zero on the visualization, I made these 
                lines more transparent so as not to overwhelm the image. This way, the 
                volume of patients who didn't have a successful outcome could still be 
                visible and the point where treatment failed is clear.
                
                

                The user can select which group(s) they would like to visualize: 
                patients with a male partner, a female partner, or no partner. In 
                the following image, only patients with a male partner are visualized.
                
                

                I added the ability to select an individual patient by shift-clicking on
                a data-point or randomly selecting one with a button click. This grays
                out the other data and makes the patient's line thicker.
                
                I wanted to display data about the patients in a paragraph of plain text rather 
                than over each stage so that a more wholistic journey was presented. When
                a patient is selected, a paragraph listing how many eggs they had mixed, embryos
                created and transferred, foetal sacs with a pulse were found, and the 
                final outcomes of IVF. It also notes if they have had previous rounds of IVF.
                
                
 
                 
            
 
            
             
            Code