I have an index named firedrill. All documents in that index have a field named DrillSteps:
"DrillSteps": [
{
"Errors": [
"Oven door opened before fuses were turned off."
],
"NumErrors": 1,
"Name": "Fuses turned off."
},
{
"Errors": [
"Fuses turned on again before oven was safe",
"Door was not closed again fast enough"
],
"NumErrors": 2,
"Name": "Oven door closed again."
},
{
"Errors": [
"Oven opened before protective gear worn."
],
"NumErrors": 1,
"Name": "Protective gear worn."
},
{
"Errors": [],
"NumErrors": 0,
"Name": "Oven smoke died out."
},
{
"Errors": [],
"NumErrors": 0,
"Name": "Burnt meal disposed of."
}
]
All you can see, each step has a name, and a list of errors committed by the student when going through that particular step. Each drill step has a specific list of possible errors. E.g. the error "Oven opened before protective gear worn" is specific to the step "Protective gear worn".
I would like to create a graph, where the y-axis shows drill errors grouped by drill steps. Importantly, drill errors that are not possible values for a drill step are not shown. The x-axis should display the error count.
Currently, this is how I have defined my graph:
And this is the output:
The count on the x-axis refers to the document count.
As you can see, this is not what I want. I am obviously doing something wrong, so I would appreciate receiving feedback on what changes I must make (either to how I defined my graph or to how I designed the data schema) in order to display my desired graph.
If anything is unclear, please let me know and I will rephrase and/or provide more information.