clean isp rd1

isp_eda.qmd

@@ -76,8 +76,8 @@ isp <- read.csv("data/isp.csv",
                                has_satel = "logical",
                                array_agg = "character"))
 
-isp[["cnt_locations"]] <- isp[["cnt_locations"]]                               
-isp[["cnt_locations_services"]] <- isp[["cnt_locations_services"]]  
+isp[["cnt_locations"]] <- as.numeric(isp[["cnt_locations"]])                               
+isp[["cnt_locations_services"]] <- as.numeric(isp[["cnt_locations_services"]])  
 
 table_with_options(isp)
 ``` 
@@ -95,7 +95,7 @@ sprintf("Number of unique brand name pre cleaning: %s", length(unique(isp[["bran
 if we remove all capitalization:
 
 ```{r}
-sprintf("Number of unique brand name pre cleaning: %s", length(unique(tolower(isp[["brand_name"]]))))
+sprintf("Number of unique brand name w/o caps: %s", length(unique(tolower(isp[["brand_name"]]))))
 ```
 
 We can filter, on this release, 21 cases for brand name.
@@ -106,5 +106,36 @@ We can filter, on this release, 21 cases for brand name.
 - provider_id can be wrong 
 - brand name can be different 
 
+One case: 
 
+0003738655	130432 	"EATEL Corp."	83537	86548	true	true	true	false	false	{LA}
+0009873712	131103 	"EATEL Corp."	34494	34497	false	true	true	false	false	{LA}
 
+Other case: 
+
+0002626984	130008 	Acentek	47	47	true	false	true	false	false	{MN}
+0002626984	130008 	ACENTEK	1395	1395	true	false	false	false	false	{MN}
+0002645927	130008 	Acentek	19521	26636 true	false	true	true	false	{IA,MN}
+
+
+### Rules for problems: 
+
+- less than 2 locations
+
+```{r}
+#| label: less than 2 locations
+isp[["quality"]] <- "Good"
+isp[isp[["cnt_locations"]] < 2, "quality"] <- "few locations"
+```
+
+- More than one frn for a provider_id "
+
+```{r}
+#| label: more than one frn for a provider
+#isp[isp[["cnt_locations"]] < 2, "quality"] <- "few locations"
+```
+
+
+- same frn but more than one provider_id
+
+- frn and/or provider_id with more than one row