Merge pull request #48 from cgostic/master

Final updates to visualizations, edits on report

scripts/03_EDA.py

@@ -96,18 +96,18 @@ def main(file_path_raw, file_path_pro, accepted_plates_csv, rejected_plates_csv,
     # Add column with length of n-gram
     counts['ng_length'] = counts['ngrams'].str.len()
 
-    n_g_len_chart = (alt.Chart(counts.query('ng_length%2 == 0')).mark_bar().encode(
+    n_g_len_chart = (alt.Chart(counts.query('ng_length%2 == 0')).mark_bar(color = 'darkorange').encode(
             x = alt.X('counts:O', 
-                title = "Frequency of given ngram in training data"),
+                title = "N-gram freq. in training data", axis=alt.Axis(labelAngle=0)),
                 #scale=alt.Scale(domain = (0,89))),
-            y = alt.Y("count()", scale=alt.Scale(type='log', base=10), title = 'Ngrams with X Freq.'),
+            y = alt.Y("count()", scale=alt.Scale(type='log', base=10), title = 'N-grams with X freq.'),
             facet = alt.Facet('ng_length:N', title = 'n-gram length')
-        ).configure_axis(labelFontSize=15,titleFontSize=10
+        ).configure_axis(labelFontSize=10,titleFontSize=12
         ).configure_header(labelFontSize=15
         ).configure_title(fontSize=20, anchor = 'middle'
         ).properties(title = "Counts of n-gram frequency by length", 
-                    width = 400, 
-                    height = 80, 
+                    width = 175, 
+                    height = 125, 
                    # columns = 1,
                     background = 'white'))
 

scripts/04_data_model.py

@@ -108,7 +108,7 @@ def main(file_path_read, filename_x_train, filename_x_validate, filename_x_test,
     line = alt.Chart(tr_v_plot_df).mark_line().encode(
         x = alt.X('Number of Features:Q'),
         y = alt.Y('value:Q', title = 'Accuracy score'),
-        color = alt.Color('variable:N', legend = alt.Legend(title = "", orient = 'bottom'))
+        color = alt.Color('variable:N', legend = alt.Legend(title = ""))
         )
 
     point = alt.Chart(tr_v_plot_df).mark_point().encode(
@@ -127,8 +127,13 @@ def main(file_path_read, filename_x_train, filename_x_validate, filename_x_test,
     (line + point + text).configure_axis(labelFontSize=15,titleFontSize=15
         ).configure_header(labelFontSize=15
         ).configure_title(fontSize=20, anchor = 'middle'
+        ).configure_legend(
+            orient = 'none',
+            fillColor = 'white', 
+            legendX = 475,
+            legendY = 250
         ).properties(width = 700,
-         background = 'white', title = 'Training and Validation Error by n-gram Range and Number of Features'
+         background = 'white', title = 'CV Training and Validation Error by Number of Features'
          ).save(file_path_write + 'train_val_error.png', scale_factor = 2)
 
     # Train model with chosen n-gram length range (2,2)