added updated md docs for FE

Signed-off-by: Duncan Ragsdale <[email protected]>

frontend/public/assets/1/SubmissionInstructions.md

@@ -6,7 +6,17 @@ For each AC power stream analyzed, there are two sets of data:
 1) The associated time series data for the AC power, in kW, and an index for timezone-aware datetime.
 2) The time shift validation file, which gives the associated time shift amount on a daily basis. A daily timezone-unaware timestamp is used as the associated index. 
 
-An example time series and its associated time shift data are shown in Figures XX and XX, respectively. An example dataset is provided for reference here.
+An example time series and its associated time shift data are shown in Figures 1 and 2, respectively. An example dataset is provided for reference here.
+
+![image info](./time_series_data.PNG)
+
+*Figure 1: Snapshot of the structure of the time series data analyzed for time shifts.*
+
+![image info](./time_series_validation_data.PNG)
+
+*Figure 2: Snapshot of the structure of the ground truth/validation data that time shift results are compared against.*
+
+Download an example data set here.
 
 ### File Submission Requirements
 
@@ -48,5 +58,20 @@ def detect_time_shifts(time_series,
 
 ### Analysis Results
 
+An automated results report containing the following metrics is returned following successful analysis:
+
+1) Mean average error (MAE) across all of the test cases. MAE is calculated at the time stamp level. The predicted time shift at each time stamp is compared to the ground-truth time shift, and error across all time shifts is calculated.
+2) Average run time per instance (in this case, by data stream), calculated in seconds
+3) Associated graphics, including error distributions color-coded by issue type (DST, partial DST, random shifts) and frequency of the time series evaluated (1 minute, 5 minute, 15 minute, etc.), and a distribution of instance run times. Some example graphic outputs are shown in Figures 3, 4, and 5.
+
+![image info](./histogram-mae-issue.png)
+
+*Figure 3: Error distribution of algorithm results, color-coded by issue type.*
+
+![image info](./histogram-mae-data-freq.png)
+
+*Figure 4: Error distribution of algorithm results, color-coded by time frequency of the input time series.*
 
+![image info](./run_time_dist.png)
 
+*Figure 5: Distribution of algorithm run times on individual time series data streams, in seconds.*