+
+This graphic is pretty tricky to make sense of because there are so many data points and drug names that are all jumbled together. So, let's take a look at the DBSCAN clustering result instead:
+
+
+``` r
+clusters_list <- tsne_df |>
+ dplyr::group_by(Cluster) |>
+ dplyr::summarize(Drugs = paste(Label, collapse = " | ")) |>
+ dplyr::arrange(Cluster)
+
+ # Print each cluster and the associated drug names
+ for (i in 1:nrow(clusters_list)) {
+ cat("Cluster", clusters_list$Cluster[i], ":\n")
+ cat(clusters_list$Drugs[i], "\n\n")
+ }
+```
+
+```
+## Cluster 1 :
+## acetylsalicylic acid and corticosteroids | aluminium preparations | aminophylline | amphotericin B | betaine hydrochloride | bexagliflozin | bupivacaine and meloxicam | calcium lactate | captopril | carumonam | casopitant | ceftibuten | chymopapain | clotiazepam | cyanocobalamin | desonide and antiseptics | dexamethasone and antiinfectives | diisopromine | epinastine | eptifibatide | ferric acetyl transferrin | fluciclovine (18F) | fluorouracil, combinations | flutrimazole | folic acid | gatifloxacin | gefarnate, combinations with psycholeptics | histapyrrodine, combinations | idrocilamide | iodine iofetamine (123I) | istradefylline | Kwas acetylosalicylowy i kortykosteroidy | Kwas foliowy | lactulose | levodopa | levonorgestrel | magnesium carbonate | methaqualone | moexipril and diuretics | narcobarbital | nimetazepam | pegloticase | phenylephrine | pipotiazine | plerixafor | reteplase | rivastigmine | roxithromycin | salsalate | sorbitol | succinimide | teneligliptin | theophylline, combinations excl. psycholeptics | tiemonium iodide and analgesics | timolol, thiazides and other diuretics | tolperisone | tramadol | tretoquinol | trypsin, combinations | ursodoxicoltaurine | Węglan magnezu | zidovudine
+##
+## Cluster 2 :
+## antazoline | azacitidine | benzocaine | taurolidine
+##
+## Cluster 3 :
+## artesunate and amdoiaquine | artesunate and amodiaquine
+##
+## Cluster 4 :
+## benazepril and amlodipine | benazepril nd amlodipine
+##
+## Cluster 5 :
+## benzoly peroxide | benzoyl peroxide
+##
+## Cluster 6 :
+## betametahsone | betamethasone
+##
+## Cluster 7 :
+## betaxolol, combinations | betaxolol, cominations
+##
+## Cluster 8 :
+## biperiden | pirprofen
+##
+## Cluster 9 :
+## buspirone | pethidine | prednisone
+##
+## Cluster 10 :
+## calcium lactate gluconate | calcium lactate guconate
+##
+## Cluster 11 :
+## cefapirin | eosin | mecasermin | prazosin
+##
+## Cluster 12 :
+## dfluprednate | difluprednate
+##
+## Cluster 13 :
+## digitalis leaves | digitalis lleaves
+##
+## Cluster 14 :
+## eplonteresn | eplontersen
+##
+## Cluster 15 :
+## flumetasone | fulmetasone
+##
+## Cluster 16 :
+## fostemsavir | fostemsvir
+##
+## Cluster 17 :
+## Hyerici herba | Hyperici herba
+##
+## Cluster 18 :
+## indometacin, combinations | indometacinn, combinations
+##
+## Cluster 19 :
+## isoprenaline | perphenazine
+##
+## Cluster 20 :
+## kanamycin | lincomycin | lnicomycin | rifamycin
+##
+## Cluster 21 :
+## megestrol and estorgen | megestrol and estrogen
+##
+## Cluster 22 :
+## menadione | menadionee
+##
+## Cluster 23 :
+## micafungin | ticarcillin
+##
+## Cluster 24 :
+## nebivolol ad amlodipine | nebivolol and amlodipine
+##
+## Cluster 25 :
+## odevixibat | oedvixibat
+##
+## Cluster 26 :
+## potassium citrate | pottassium citrate
+##
+## Cluster 27 :
+## remoxiipride | remoxipride
+##
+## Cluster 28 :
+## srteptokinase | streptokinase
+##
+## Cluster 29 :
+## technetium (99mTc) perechnetate | technetium (99mTc) pertechnetate
+```
+
+For instance, cluster 1 includes a wide variety of drugs, such as acetylsalicylic acid and corticosteroids, aminophylline, bupivacaine, and meloxicam.
+
+Clusters like cluster 5, which includes benzoyl peroxide and its misspelled variant benzoly peroxide, show how useful Levenshtein distance can be for identifying typographical errors. Similarly, cluster 6 brings together different spellings of betamethasone.
+
+This clustering analysis helps us understand the relationships among drug names and the potential issues that similar or identical names can cause. This isn't perfect, though. We still need to check if there are any other similar duplicates (take a look at the first cluster). Also, the Levenshtein metric is designed for editing, so it's not ideal for texts in different languages (see "Kwas foliowy" and "folic acid"). But it can definitely streamline our work and automate the process for future reports, and it provides some interesting insights!
+
+
+# Final Thoughts: Con-meds data in EDC - how to make it clean and meaty?
+
+It's really important to make sure that the data on the drugs people are taking at the same time as the ones being tested (concomitant medications, or "con-meds") is accurate and complete in the systems we use to capture data from clinical trials. To make this happen, we need to automate data entry tasks. Instead of relying on researchers to manually enter drug data, integrating the electronic case report form (eCRF) with a comprehensive drug dictionary, like the WHO Anatomical Therapeutic Chemical (ATC) classification, can make the process a lot more efficient. This integration would make it simple for users to enter or select drug names from a standard list, which would help to avoid errors caused by spelling variations or outdated names. By using these automated solutions, we can make sure that our con-meds data is cleaner and more reliable, which will improve the overall quality of clinical research.
diff --git a/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-10-1.png b/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-10-1.png
new file mode 100644
index 0000000..349a1ab
Binary files /dev/null and b/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-10-1.png differ
diff --git a/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-11-1.png b/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-11-1.png
new file mode 100644
index 0000000..be0e69b
Binary files /dev/null and b/content/post/2024-10-01-same-same-but-different-how-advanced-data-science-techniques-help-us-validate-drug-names/index_files/figure-html/unnamed-chunk-11-1.png differ
diff --git a/content/search.md b/content/search.md
new file mode 100644
index 0000000..70d339b
--- /dev/null
+++ b/content/search.md
@@ -0,0 +1,6 @@
+---
+title: "Search" # in any language you want
+layout: "search" # necessary for search
+summary: "search"
+placeholder: "What are you looking for?"
+---
diff --git a/layouts/_default/single.html b/layouts/_default/single.html
new file mode 100644
index 0000000..5c0694c
--- /dev/null
+++ b/layouts/_default/single.html
@@ -0,0 +1,75 @@
+{{- define "main" }}
+
++ {{ .Title }} + {{- if .Draft }} + + + + {{- end }} +
+ {{- if .Description }} +
+ {{ .Description }}
+
+ {{- end }}
+ {{- if not (.Param "hideMeta") }}
+
+ {{- end }}
+
+ {{- if not (.Param "disableAnchoredHeadings") }}
+ {{- partial "anchored_headings.html" .Content -}}
+ {{- else }}{{ .Content }}{{ end }}
+
+ {{- end }}
+
+ + + +
+
+
+ {{- if (.Param "comments") }}
+ {{- partial "comments.html" . }}
+ {{- end }}
+
+
+
+ 
+
+ Let’s solve real-world problems together — contact us at office@ttsi.com.pl to explore how data science can drive your success.
+
+ 
+