Project description:

For an article in AthensLive, Sotiris Sideris collected data on properties up for bidding through eauction.gr, Greece’s official website for auctioning real estate seized from over-indebted borrowers. The dataset includes 45,918 lots listed between mid-November 2017 (when the website launched) and September 1, 2019. For each lot, the dataset specifies the auction date, property characteristics, starting bid, total debt, debtor, hastener pursuing the auction, links to additional documentation, and more.

Impact reached:

Picture the ranks of people losing their homes or small businesses in auctions because they couldn’t pay back their debt to the bank: families with children sofa-surfing, at least for a while, with friends and relatives; homelessness; the social stigmatization associated with unpaid debt which can lead to an increase in mental health problems.

 

Before starting our research into private property auctions, we didn’t know how many private properties have gone under the hammer, how often or for how much debt. Most importantly, we didn’t necessarily understand the disorienting reality for so many of living like this.

As part of our research, we talked to people who have to cope with such a dire situation and we followed activists, who have been trying to block the auctions of primary residencies, on the ground. Also, we presented the findings of our research at data conferences and academic events both in Greece and abroad. 

There were many insights throughout the process but the stand-out was trust. The media market in Greece is characterised by a lack of trust and poor use of traditional media (including a weak newspaper market) alongside some of the highest use of social media and digital-born outlets. At AthensLive we create a climate where journalists can regain public trust and the public’s interests can be represented in journalism again through democratic participation and transparency. 

Techniques/technologies used:

To collect data on properties up for bidding we scrapped eauction.gr using the Beautiful Soup library and Selenium framework in Python. The first round of scraping gave us basic information about each auction but most importantly the URL of each entry. Following those URLs we looped through all 45,918 entries in a second round of scraping in order to collect relevant data for each lot, such as the auction date, property characteristics, starting bid, total debt, debtor, hastener pursuing the auction, links to additional documentation, and more. Then, in the final round of scraping, we looped through every entry again, this time to download the court decision for each unique online auction in PDF and Word Doc format, in Greek language. Among the individual debtors several have multiple properties against their names. On August 31, at least 30,134 individuals had more than one property listed. Some of the individuals clearly held large property portfolios. In one instance, a single person is named against 87 lots. Therefore, there are 22,119 items of documentation in our database instead of 45,918 – which is the exact number of electronic auctions within our study period.

After we collected all the relevant information from eauction.gr, we converted our lists of results into a dataframe in the Pandas library in order to start the cleaning and then the analysis. We cleaned our dataframe using regex and then we analysed it in Pandas. When it comes to the visualization of our results, we used the Matplotlib plotting library for Python and then exported our graphs to edit them in Adobe Illustrator.

What was the hardest part of this project?

To better understand how this platform accelerates the processing of auctions, we wanted to look in some detail at every property that has appeared online. However, unsurprisingly, no data is available online; nothing from the Notary Association website; nothing from Athens Bar Association website. The only way to collect this data was to loop through every property on eauction.gr to collect the relevant information and put it into a dataset of 45,918 lots that can then be analysed. Challenging, but totally worthwhile.

We managed to map 26,976 properties by regional unit. This number, however, represents 58.7% of all the properties in question as real estate type and location information are not available for auctions with posting date before September 24, 2018. This kind of information, simply put the exact address of the real estate, is only available in the documentation of each auction in our database. Mapping every single auctioned property would allow us to find hot sports of gentrification, eviction, and displacement.

When it comes to bidders, we don’t really know who bought what and for how much. That’s how financialization works; it serves to allow anonymity for the rich. On the other hand, land – and the devastation caused to the lives of individuals and communities who have lost it – is personal and the resulting wave of displacement can lead to an increase in mental health problems.

If we want housing to be a public good and not a vehicle for investment, we need more democratic control, not less. We must increase the visibility of the problem and advocate for those who are at risk of losing their homes, knowing that the struggle against the seizure of private property is a struggle for human rights.

What can others learn from this project?

Undoubtedly, the online auction platform has dramatically expedited the rate of auctions. Essentially, though, it has de-linked real estate and place by making the intrinsically local and fixed nature of real estate into something liquid and therefore tradable on global financial marketplaces. “By using the online auction service, you can participate in online auctions without your physical presence using a computer.” That’s the first thing you read when you enter the platform.

In terms of prices, there is great disparity in the starting bids for individual lots, with €52,000 being the median. At the top of the scale are the production facility of the solar power generation company HelioSphera in Tripoli, Arcadia (€37,449,000); the facilities of IRIS printing company in Koropi, Attica, one of the largest in Southeast Europe (€35,183,140). Also at the top is ‘the most expensive house in Athens,’ a six-storey luxury residential tower minutes away from Syntagma Square (€35,000,000), whose owner, arms dealer Konstantinos Dafermos, was arrested in 2015 for illegal payments given to secure a contract to supply Kornet anti-tank missile launchers to the Greek armed forces. At the bottom of the price range are storage rooms for €900, cars for €300, individual parking spaces, even coffee tables.

Journalism in Greece has been in a dismal state for years. Overall, the project is a contribution to the growth of independent journalism projects in a multicultural and international community.

Project links:

medium.com/athenslivegr/whose-home-is-this-f3b45d878b0b

Open dataset and explanatory notes [tinyurl.com/s9t454v]

Greek version of the article [tinyurl.com/rdzcla9]

Video of activists blocking an auction in Athens [tinyurl.com/tyhj2fr]

Reference and video interview by CNN Greece [tinyurl.com/ughe65m]