Update readme 📊📉

R/preprocess.R

@@ -96,6 +96,7 @@ preprocess_bot.data.table <- function(x, batch_size = 100, ...) {
   if (is.null(batch_size) || isFALSE(batch_size) || length(uid) <= batch_size) {
     x <- preprocess_bot_group(x)
     if (!is_ids(ogusrs)) {
+      screen_name <- NULL
       x <- x[match(tolower(ogusrs), tolower(x[, screen_name])), ]
     } else {
       x <- x[match(ogusrs, x[, user_id]), ]
@@ -278,6 +279,9 @@ preprocess_bot_group <- function(data) {
   usr_actyr <- NULL
   tweets <- NULL
 
+  dtime <- NULL
+  varname <- NULL
+
   ##----------------------------------------------------------------------------##
   ##                         DTIME (TIME BETWEEN TWEETS)                        ##
   ##----------------------------------------------------------------------------##
@@ -287,7 +291,8 @@ preprocess_bot_group <- function(data) {
   m <- m[, .(dtime = c(NA_real_, abs(as.numeric(diff(created_at), "mins")))), by = user_id][
     order(user_id, dtime), .(dtime, varname = paste0("dtime", seq_len(.N))), by = user_id]
   ## create complete version of dtimes (with missing values)
-  mna <- data.table::data.table(user_id = unique(m[, user_id]))[, .(dtime = NA_real_, varname = paste0("dtime", 1:200)), by = user_id]
+  mna <- data.table::data.table(user_id = unique(m[, user_id]))[,
+    .(dtime = NA_real_, varname = paste0("dtime", 1:200)), by = user_id]
   ## merge the two–removing duplicated rows from the NA dataset
   m <- rbind(m, mna)[!duplicated(data.table::data.table(user_id, varname)), ]
   ## convert from long to wide for each user

README.Rmd

@@ -176,11 +176,14 @@ The most influential features in the classifier
 
 ```{r, include=FALSE, eval=FALSE}
 # idsbot <- c("user_id", "screen_name", "bot")
+# mod <- tweetbotornot2:::prep_xgb_model()
+# mod$feature_names <- tweetbotornot2:::tweetbotornot_xgb_model_feature_names
 # xgboost::xgb.ggplot.importance(xgboost::xgb.importance(
-#   model = tweetbotornot2:::tweetbotornot_xgb_model,
-#   trees = 1:1000), measure = "Gain", top_n = 28) +
+#   model = mod,
+#   trees = seq_len(tweetbotornot2:::tweetbotornot_xgb_model_best_ntreelimit)),
+#   measure = "Gain", top_n = 20) +
 #   ggplot2::scale_fill_viridis_d(begin= 0.05, end = 0.9) +
-#   dataviz::theme_mwk(12, "Avenir Next LT Pro") +
+#   dataviz::theme_mwk(14) +
 #   #ggplot2::coord_cartesian(ylim = c(0, 0.3)) +
 #   #ggplot2::coord_flip(ylim = c(0, .12)) +
 #   ggplot2::ggsave("man/figures/README-import.png", width = 9, height = 8,
@@ -195,15 +198,17 @@ How features contributed to predictions in the original training data:
 
 ```{r, include=FALSE, eval=FALSE}
 # .d <- tfse::read_RDS("../twbt/data-final-munged.rds")
+# mod <- tweetbotornot2:::prep_xgb_model()
+# mod$feature_names <- tweetbotornot2:::tweetbotornot_xgb_model_feature_names
 # 
 # png("man/figures/README-shap.png", width = 9, height = 8, units = "in", res = 312)
 # par(tcl = -0.175, bty = "n", xaxt = "s", yaxt = "s", col = "#aaaaaa")
 # cols <- viridis::viridis_pal(begin = 0.1, end = 0.9)(2)
 # suppressWarnings(
 #   xgboost::xgb.plot.shap(
 #     data = as.matrix(.d[, -(1:3)]),
-#     trees = 1:150,
-#     model = tweetbotornot2:::tweetbotornot_xgb_model,
+#     trees = seq_len(mod$best_ntreelimit),
+#     model = mod,
 #     top_n = 36, n_col = 6,
 #     discrete_jitter = 0.15, span_loess = 0.35, col_loess = cols[1],
 #     col = cols[2],