30 Haziran 2018 Cumartesi

Otomatik olarak R ile günlük analiz yapıp git ile güncelleyin

Bu kodları çok farklı sayıda blog ve yazıdan derledim :)

Hedefim şu. Elimde bir R Markdown ya da R Notebook dökümanı var. Bunu belirli periyotlarla çalıştırıp güncel analizi git komutları ile github sayfama yüklemek istiyorum.

Böylece belirli bir API'den çektiğim veriyi belirli periyotlarla analiz edebileceğim.
Bu şekilde PubMed verisi üzerinde güncel analizlerimi oluşturmaya başladım.

Örnek otomatik analizler: 

Örnek PubMed analizleri:

1) R-project kodunu belirli periyotlarla çalıştırmak için RStudio'nun eklentilerini kullanıyorum:

Böylece aşağıdaki R dosyasını otomatik olarak belirli periyotlarla çalıştırmak mümkün.

2) Otomatik çalışan bu R kodunu kullanarak analizimi içeren R Markdown dosyasını çalıştırıp html çıktısını oluşturuyorum:

rmarkdown::render(input = "ScheduledAnalysis.Rmd",
output_format = "html_notebook",
output_file = "docs/ScheduledAnalysis.html",
quiet = TRUE)

3) Aynı R kodu içinde rstudioapi paketi ile de comand line, bash, terminal komutlarını da otomatik olarak çalıştırmak mümkün. Böylece git push da aynı anda yapılabiliyor.

CommitMessage <- paste("updated on ", Sys.time(), sep = "")
gitCommand <- paste("git add . \n git commit --message '",
"' \n git push origin master \n", sep = "")
gitTerm <- rstudioapi::terminalCreate(show = FALSE)
rstudioapi::terminalSend(gitTerm, gitCommand)

23 Haziran 2018 Cumartesi

ICD-O sınıflama sisteminin Türkiye'de kanser kayıt sistemi için kullanılması

ICD-O sınıflama sisteminin Türkiye'de kanser kayıt sistemi için kullanılması yönünde uzun süredir devam eden çalışmalar var.
Aşağıda bu konuyla ilgili olarak ResearchGate'de ( yapılmış yazışmaları ve bu konu ile ilgili yayını Dr.Canda'nın ( da izni ile paylaşıyorum.
Ben uzun süredir pratikte çalışmadığım için güncel durumdan habersizim, sadece bibliyografik çalışma hobim ( nedeniyle bu konu ile ilgilenmeye devam ediyorum. Kanser istatistiklerinin düzenli tutulması, büyük bölge hastanelerinden yıllık kongrelerde düzenli güncel sayıların sunulması ve trendlerin incelenmesinin önemine hep inandım.
Özellikle büyük bölge hastanelerinde çalışan patologların bu konu ile ilgili öneri, katkı ve eleştirilerini Dr.Canda'ya ya da federasyona iletmesi bu konudaki çalışmalara destek olacaktır.
Dear Serdar Balcı,
I would like to know to your views on the following topics.
- "ICD-O 3.1 Turkish version"
- "New Histopathological Diagnostic Terms and their Codes of the Central Nervous System (CNS) in the International Classification of Diseases for Oncology (ICD-O 3.1)"
-The importance of the coding of the pathology reports according to ICD-O 3.1 for Turkey which is I'm working since 1991 up to now.
With my best wishes.
Dear Prof Canda,
Sayın Hocam,
Thank you for sharing your work on ICD codes with me. I would like to thank and congratulate you on this effort.
I am happy for this opportunity to share my ideas on this topic with you.
I have experience in SEER data and have actively participated in gathering and analyzing SEER data during my research days in Atlanta, in Emory Epidemiology Department with Dr.M.Goodman. I have also published articles using these community data. I always have dreamed to have a similar database in Turkey as well. I have not been in practice for a while, so I am not aware of the current situation. But as I used to be responsible for the pathology information system and quality control, I have experienced many problems in cancer registry process.
First and foremost, the registry is now defined as a duty for pathologists, unlike to those of other countries where registry is performed by other professionals namely "registrars". Considering the amount of workload on a practicing pathologists, this is becoming as a burden, as I believe that the codes are not easy to apply in routine practice.
Pathologists are used to diagnose with histologic codes and WHO blue books. It would be more easy for pathologists, if they are asked to name which WHO blue book they used and which histology code they gave. Any pathology information system can be adapted very easily for this. The data retrieved as histologic code and major organ/system can then easily be converted to ICD codes via computers.
I do not think that details of location (whether the kidney tumor is on right or left kidney) is not required. But the selection of kidney and appropriate "papillary carcinoma" is important. I have faced many "papillary carcinoma" or "in situ ductal carcinoma" codes mistaken among various organs, as well as "basal cell carcinoma" mistaken for other basaloid neoplasms.
In my research on SEER data of gallbladder and pancreas tumors, I found that tumors are inappropriately coded histologically. Even major subtypes are not easy to identify, not to mention confusing benign and malignancy terms. Since pathologists are now the official registrars here, at least we may have an opportunity not to make the same mistake SEER has made decades ago.
If we use histologic codes primarily than it will be much easy to see trends in diagnosis and treatments as well as it will be easy to track when the nomenclature changes, which is the hobby of pathologists :)
I would like to thank you again for asking my opinions on this topic.
Best regards.
Serdar Balcı, MD, Pathologist
Many thanks for your detailed and informative response.
-All Turkish pathologist must contribute to and do ICD-O work in the world in the scientific field.
-In terms of ICD-O's implementation, what should we do about it in Turkey.
My purpose is as follows:
a) The ICD-O system has been used worldwide since 1976.
b) Turkey has also decided to implement since 2015.
c) For Turkey, this is a very important development.
d) It is important that "ICD-O 3.1 Turkish version, 2017" is implemented in a healthy manner.
e) What should be done for this purpose?
f) What are the views of the pathologist?
g) I believe in the benefit of being a "solution-oriented debate".
Conclusions and recommendations;
1. I have been working since 1991 to use the ICD-O coding system.
2. Cooperation of the Ministry of Health and Universities is very importend.
3. Persons who believe in the importance of this matter and have knowledge should be taken into consideration.
4. For the computer coded program, the opinion of pathologists must be taken.
5. Those who have worked outside the Ministry of Health and NGO representatives should also be considered.
6. Conclusion: Loading "additional workload" into pathology specialists does not solve the problem.
7. The work should be solution focused.
8. The coded program should be open to criticism from pathologists.
9. What are the opinions of Pathologists?
Greetings, best regards
Dear Prof Canda,
With your permission I would like to copy our conversation to facebook groups and other social media, in order to have more opinions from Turkish pathologist.
Please let me know if I can copy this conversation and ask friends opinion.
Best wishes
Good morning, Greetings and good days.
Thank you for your suggestion. I hope that better ideas emerge.
(-Due to the "performance system" in hospitals, the times and "pensions" of especially idealistic Doctors are very limited.)
-My objective is also the coding of Tumor Pathology Reports according to "ICD-O 3.1 Turkish version, Ministry of Health of Turkey, 2018, Ankara".
-To encode the Tumor Pathology Reports, it is necessary to make an effort (+) together (Ministry of Health and Universities-Pathologists, Cancer Registrers and Computer Programmers).
-I have been working on this subject since the year of -1991.
-Yet, it was not widespread use of computers in Turkey in that period.
-In my assistantship years, there were some studies in the Department of Pathology of Ege University Faculty of Medicine (1972-1975).
Prof. Dr. Yavuz Aksu was working on the Index-Card system.
My Professour Yavuz Aksu made an effort to do "Index Card system" and he indicates the importance for scientific studies and for Turkey.
-I also tried to collect regional cancer cases in the Ataturk University, Cumhuriyet University and Dokuz Eylül University in the following years and published in a part of it.
-I published some of the my studies at Ataturk University in the Ankara Bulletin of Pathology (Editor Prof. Dr. Muharrem Köksal, 1976,1977).
-Nowadays computer programming systems and digital revolution are very important developments.
If we do that;
-without being too late
- as "open" minded scientists,
- working with a focus on "solving contribution"
-Turkey can also take place among the Countries that effectively implement ICD-O in the world.
Furthermore, the fact that "ICD-O Turkish version"
-It also allows Turkish to become one of the most spoken languages ​​in the world.
-If it was an alphabetical unity, other Turkish-speaking countries could have been involved in this system.
In Turkey
-cancer map and regional cancer studies
- in terms of frequency of cancer and public health problems,
- in terms of making things healthy in this area and
- It allows the Turkish doctors and Turkish cancer patients to use the same language "at the same time" with the world.
- Turkey will be get much more healthy cancer statistics
- Turkey and the Turkish scientists of international reputation increases.
I offer my respects to all colleagues who will contribute to this issue.
Greetings, sincerely yours
Prof. Dr. M. Şerefettin Canda
June 22, 2018, Güzelbahçe-İzmir

İlk paylaşım linki

9 Haziran 2018 Cumartesi

GitHub sayfamı oluşturdum

Microsoft'un satın almasından sonra bazı göçler yaşansa da ben daha yeni öğrendiğim bu mecraya yavaş yavaş yerleşiyorum. Bugün GitHub anasayfamı oluşturdum:

Biraz uğraştırsa da şu linklerdeki yönlendirmeler gayet işe yarıyor:

GitHub Pages

Jekyll is a simple, blog-aware, static site generator.

Creating and Hosting a Personal Site on GitHub

4 Haziran 2018 Pazartesi

Devridaim makinası

Devridaim oyuncakları hep ilgimi çekmiştir.

Periyodik tablo videolarının meşhur Profesörü'nün tanıttığı bu "dönergeç" gayet ilginç:

Can you solve the mystery of The Professor's Perpetual Motion machine?

Perpetual Motion Machine - Periodic Table of Videos