ENGLISH

News Deserts Spread in Turkey’s Inner Aegean, Central Anatolian and Eastern Regions

This map visualizes the gross number of Turkish news articles that cited each district of Turkey in a two-week period. We used Datawrapper for map visualizations, Tableau for data analysis, and Google Data Gif Maker for animated graphics.

Over the course of two weeks, we scraped 1,116,916 news articles from more than 6,500 Turkish news websites in which the names of 973 districts appear.

Which districts have the highest or lowest coverage? Where do we see original journalism the most and where exactly lacks it? Which districts of Turkey face the risk of emerging news deserts?

To answer these questions, let’s take a look at the three maps that we have created with the help of data scientist Emrah Yılmaz.

(This article was originally published in Turkish on Feb. 24. It is translated by Şebnem Özakıncı, and edited by Olivia Flaherty-Lovy, who both participated in Journo’s New Horizons internship program in the summer of 2021. You can find our original dataset here.)

1. The district with the highest news coverage in Turkey: Fatih, Istanbul

We collected the data between 7-20 February 2021. The map above shows the total number of news articles per district. Apart from the metropolises, it is clear that we get much less news from the districts away from the coastal regions and borders.

Fatih, Istanbul takes the lead with 18,940 news articles in two weeks, according to our data. Istanbul, Bursa, and Kocaeli districts are among the top ten in terms of how many articles cited them.

The first district to rank highly, from a region other than Marmara, Mediterranean, or Aegean regions, is from the Black Sea region; Ereğli, Zonguldak with 5,434 news articles.

The distribution of the districts in the top-hundred ranking per region is as shown below (at least 150 news articles a day coming from these districts):

  • Only 8 locations from Central Anatolia, 6 locations from Southeastern Anatolia, and 5 locations from Eastern Anatolia can make it among the top-hundred districts with the highest number of news articles citing them. (The total number exceeds 100, due to some districts that are tied in some ranks.)
  • The first district to enter the top-hundred ranking from South East is Şahinbey, Gaziantep (ranking 52nd with 4,091 news articles).
  • The first district to enter the top-hundred ranking from Eastern Anatolia is Yüksekova, Hakkari (ranking 66th with 3,467 news articles).

2. The lowest numbers of news articles come from Central Aegean and rural Anatolian districts

The majority of districts with the lowest news coverage consist of Central Anatolian districts (except for the capital; Ankara) and Aegean districts (except for coastal areas).

Districts with less than three news articles per day are:

  • Karahallı, Ulubey, Sivaslı from Uşak;
  • Evciler, Hocalar and Kızılören from Afyon;
  • Eldivan and Yapraklı from Çankırı;
  • Sarıyahşi and Ağaçören from Aksaray;
  • Altunhisar from Niğde,
  • Hasanbeyli from Osmaniye,
  • İnhisar from Bilecik,
  • and Günyüzü from Eskişehir.

Distribution per region is as shown below:

If not from their hometown itself, it should be important for a citizen to read news from at least a neighboring district. For instance, a resident of Günyüzü, Eskişehir is usually not able to read about her hometown in news frequently. However, she can still consume news about the other districts of Eskişehir province, which are subject to news articles frequently.

The current situation in the north of Eastern Anatolia is worse, though. This is because the districts with less news coverage spread to a very large area in this region. That is to say; the readers, who already can’t read about their own district, do not have the chance to read about the neighboring districts, either. They are often forced to consume news about other districts and provinces that are far away.

3. News coverage per capita: The picture gets better in Western Black Sea but worse in Southeastern Anatolia

When we consider the populations of the districts, we see that the number of news articles per capita is also not that high in metropolises.

Districts with the highest number of news articles in proportion to their populations are Pülümür, Tunceli; Bozcaada, Çanakkale; and Dumlupınar, Kütahya.

Districts with the lowest number of news articles in proportion to their populations are Şereflikoçhisar, Ankara; Karahallı, Uşak; and Hocalar, Afyon.

Number of news articles per capita

From this standpoint, the situation does not seem so bad in the Western Black Sea region and the area encompassing the Tunceli-Sivas-Elazığ triangle.

Hakkari and Van take the leading position in the number of news articles per capita in Eastern Anatolia. To the north of the region, however, news deserts keep emerging.

When we consider population, we encounter similar news deserts spreading from the inner areas toward the coastal areas of the Aegean region.

4. 85% of news articles are duplicates: Regions that seem like oases are actually deserts of original reporting

We have evaluated and counted each one of the published news articles for the first two maps. For example, if a news agency article titled “Heavy snow in Başakşehir blocks roads” is published on 100 websites that day, we counted each, leading us to 100 news articles registered for that district that day.

That is why plurality in the numbers of news articles does not necessarily show an abundance of original reporting. It is an undeniable truth that thousands of Turkish news websites copy and paste articles from other sources, especially from news agencies like Anadolu Agency, Demirören News Agency, and İhlas News Agency.

There isn’t yet an AI technology to evaluate news in terms of its quality or contribution to the public interest. But the system developed by data scientist Emrah Yılmaz enabled us to create an “originality score” for each district by detecting which news articles were created before the others and by analyzing the linguistic differences between varying versions of the texts.

For two weeks, there were approximately 1,147 news articles per district. Only 176 of them were original news articles. That is to say, according to the algorithm we used, 85% of the news articles published in Turkish digital media are not original.

Moving from these scores, we have evaluated the number of original news articles coming from the districts. This analysis led us to this map:

As the map shows, Pülümür, Tunceli also has the leading position in the number of original news articles per capita. Pülümür is followed by Evren, Ankara; Bozcaada, Çanakkale; Tuzlukçu, Konya; Karaburun, İzmir and Kofçaz, Kırklareli.

Şereflikoçhisar, Ankara is one of the districts that have the lowest numbers of original news articles per capita. Şereflikoçhisar is followed by Selendi, Manisa; Sinanpaşa, Afyon; Kayapınar, Diyarbakır; Şefaatli, Yozgat and Şaphane, Kütahya.

Among all the districts that were cited by more than 100 news articles in two weeks, the districts that have the highest rates of original news articles were Ulubey, Ordu (60%); Durağan, Sinop (50%) and Beydağ, İzmir (46%).

Districts that have the lowest rates of original news articles are Armutlu, Yalova (5.8%); Özvatan, Kayseri (6.1%) and Bozcaada, Çanakkale (6.2%).

In certain districts, even though the gross number of news articles is quite high, the number of original news articles per capita is extremely low. An outstanding example of this is the Şehitkamil district of Gaziantep province that was mentioned in 3,284 news articles for two weeks but had an originality rate of 12%.

This is one of the most shocking revelations of this research; regions that seem like an "oasis" in the gross number of news articles are actually "deserts" in original journalism. Or the other way around…

For instance, it can be seen from Map 1 that many news articles are coming from Southeastern districts. But as can be seen from Map 3, a very low percentage of those news articles are original.

The same issue is valid for metropolises, the majority of the Aegean region and south of the Konya province. The number of news articles is high in these areas but the rate of original news articles per capita is low.

The contrary goes for the majority of the Western Black Sea and Northeastern Anatolia and also the Tunceli-Elazığ-Sivas triangle. They seem to have fewer news articles but they have a higher rate of originality.

What does ‘News Desert’ mean?

News desert is a concept that has been widely discussed with the emerging digitalization since the early 2000s, especially in the American media. Every day, local newspapers diminish, and media outlets employ fewer reporters. Regions that do not have any daily newspapers anymore are called news deserts. In the US, around 2,000 districts out of 3,143 do not have any daily newspapers.

In our research, we have used the term "new desert" in its broadest sense. As important as it is to see whether a district has newspapers, the question of how much news coverage —including online news sites— a district has also should be kept in mind.

The most important issue, rather than the numbers, is of course how influential and beneficial a news article is for the public interest.

We hope that this research has humbly contributed to a subject area with hardly any data about digital journalism in Turkey.

Two ending notes on the collected data:

  • Names of some districts in Turkey have other common connotations. For example, in many of the news articles from Balıkesir that include the word "Marmara", the use of the word may refer to the geographical region or the sea, and not the district of Marmara. Since we did not conduct any semantic analysis, we have chosen to normalize the numbers coming from for those few districts with such tricky names with the average of their neighboring districts. We have applied a similar method for some "central" districts that hardly have any news coverage and projected the average of the province in question to the central district.
  • Though it creates the basis of journalism, the Five Ws are often overlooked in Turkish digital media. Especially editors in metropolises use the names of provinces in news articles but don’t mention districts. Sometimes, both remain unmentioned. Referring to the situation above, it should be noted that we have only considered news articles that mention both provinces and districts.

Emre Kızılkaya

İstanbul Üniversitesi Siyasal Bilgiler Fakültesi'nde Uluslararası İlişkiler lisansının ardından Marmara Üniversitesi'nde Gazetecilik yüksek lisansını tamamladı. 2003-2019 yılları arasında Hürriyet gazetesinde editör, dış haberler şefi ve dijital içerik koordinatörü; Hürriyet Daily News'da idari editör görevlerinde bulundu. İstifasının ardından 2019'da Harvard Üniversitesi'nde Knight Vakfı ve Nieman Vakfı'nın ortak bursuyla misafir araştırmacı olarak medyada sürdürülebilirlik konusunda çalıştı. Viyana merkezli Uluslararası Basın Enstitüsü'nün (IPI) Başkan Yardımcılığını ve IPI Türkiye Ulusal Komitesi Başkanlığını yaptı. Journo'da proje editörü ve Harvard Kennedy School bünyesindeki Carr Center'da misafir araştırmacı olarak görevini sürdürüyor.

Journo Newsletter