Monday, 14 May 2018

Week 6 [14.05 - 20.05.2018] Social media and mental health predictions

I've found article about analysis of social media.

http://www.bbc.com/future/story/20180201-how-your-social-media-betrays-your-mood

Please, share your opinion about questions:

1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
2) Will the access of researchers to social media data be limited?
3) Do you see other threats related to social media data mining? 

30 comments:

  1. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    I don’t think in 100% but in majority yes. Firstly people very often share only part of life, mostly the happy part or the one that is supposed to make other jealous. Some people use social media like me - not posting, liking and not being active at all, just using for communication. At the same time nothing can replace standard psychological tests.

    2) Will the access of researchers to social media data be limited?

    It all depends on the willingness (or funding) of the authorities that authorise the usage of the data. I think that access for everyone will never be possible, but specialized selected institutions can get such access.

    3) Do you see other threats related to social media data mining? 

    Personal data collections has led to a widely applied practice of data analysis and personal information trading, which can highly compromise the users’ privacy and lead to many further problems such as inaccurate 'threat rating’, too much interference in private life or unauthorized publication of private data.

    ReplyDelete
    Replies
    1. As I know because of RODO laws collecting personal data will be more difficult in future.

      Delete
  2. Thank you for taking up very important topics. I suppose a good specialist can put the diagnosis today in some cases. Some people regularly publish a lot of information. I am not an active user of social media. I do not feel the need to monitor other people or to publicize information about myself. BigDate methods offer enormous possibilities of obtaining information based on the analysis of user profiles. Sometimes the data are used for scientific purposes but for the majority, they are valuable tips in marketing activities. This situation can contribute to higher prices of insurance policies, the increase of burglaries to our homes and generally weakening our privacy with all of these consequences...

    ReplyDelete
    Replies
    1. I also don't publish anything in social media. It is possible that people will cease using it. The same situation I've observed about buing TV sets. Most of my friends don't watch TV (age 25-35).

      Delete
  3. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I think that in long future this will be possible. But when it comes to predicting health not only mental using Nueral Netowrks we need to be carefull and this future should be only available for doctors not for everybody. This is because with bad prediction without doctor bad things can happen. Imagine that my CNN discovers cancer in last phase in your USG sonograms and doctor examination and test will show you that everything is ok. What would you do without doctors opinion? Same thing is with people with mental illneses

    2) Will the access of researchers to social media data be limited?
    It depends, nowadays in Europe we have very restricted access to data from social media, but in USA you can do what ever you want. So it depends from the law.

    3) Do you see other threats related to social media data mining?
    Yes, not so long time ago somebody asked the same qestion. So one thing is that if you put a lot of information about where you actually are, many of burglar can be very happy and also pedphopiles and othes.

    ReplyDelete
    Replies
    1. I think the main problem with psychological diagnosis is that each person is different, so it is very difficult to train globaly accurate model of classification. I mean someone could be considered as mentally ill only because he is different than most of people.

      Delete
  4. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I heard about 70-120 Like. Or other activities are enough to get to know people very thoroughly. I think it will be possible in 99% :) the method does not work 100% (I think)
    2) Will the access of researchers to social media data be limited?
    I think not, but the user will decide what he makes available.
    3) Do you see other threats related to social media data mining?
    The subject is quite often discussed. I think that we have already written about it many times in this forum.

    ReplyDelete
  5. Sorry for repeating this question :) Thanks for sharing your opinion.

    ReplyDelete
  6. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    100% accuracy, no one will ever give us any topic. The machine is 200% off.

    2) Will the access of researchers to social media data be limited?

    Of course! It will be limited, only we will see to what extent. Maybe (as I wrote before) people will think of it and will not publish everything about themselves, their families and their lives.

    3) Do you see other threats related to social media data mining?

    Threats are many. The most popular are:
    - advertising spam
    - Attack on your friends: Phishing account
    - Theft of a user account without the use of phishing
    - Attack "by the way"
    and other.

    ReplyDelete
  7. Ad1
    It is important to start with the fact that no psychological diagnosis is ever made at 100%. It is important what will be investigated. Personality? Temperament? There are many possibilities in psychology. Social networking choices can be indicators. (what is liked, logging frequency, number of friends, etc.) All these indicators can help you to be predictive. On the basis of which it will be possible to estimate how a given person will behave, or what choice he/she will make. There is no big difference between the creation of such indicators and psychological tests created on the basis of factor analysis.

    Ad2 It will not be because such information is valuable.

    Ad3.
    I don't see any specific threats, the information is collected in order to sell me something better (facebook works like this). The user agrees to the registration of such data in the regulations.

    ReplyDelete
  8. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    Already now with access to such broad data that we publish to the operators of social networks ourselves, we are able to create a very thorough psychological analysis but it will never be 100% correct or at least we will not be able to be sure without actually knowing that person.

    2) Will the access to researchers be social media data be limited?

    It's hard for me to answer this question, just access to data is not dangerous and very much needed, so it should not be limited. If it is limited then it should be limited first to third parties using this data for their own profit.

    3) Do you see other threats related to social media data mining?

    Yes, of course, even if only from public data, you can take advantage of the features that they do not have in themselves - the so-called emergence. I mean that a person who publishes a lot of private information about themselves, their preferences and choices gives something more than just a set of data about themselves, but enables an easier psychological analysis of the person and thus is more susceptible to threats from the oppressor.

    ReplyDelete
  9. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I think there will be cases where diagnosis accuracy will approach 100%.
    But then what? Some research institution will refer people to therapists? Or send kind email "you have mental issues"?
    2) Will the access of researchers to social media data be limited?
    It is happening now thanks to GDPR being rolled out and increasing media coverage.
    3) Do you see other threats related to social media data mining?
    I am still surprised when - from time to time - I realize that there are people that participate in social media and care about those. Once I'll get used to it, I will came up with other threats than ones related to "hacking" activity.

    ReplyDelete
  10. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    I think that this is impossible. I think nothing 100% can not be determined This depends a lot on the information that users have invested. Is there information, not a fake one? I think that psychological tests would also help.

    2) Will the access of researchers to social media data be limited?
    It is happening now thanks to GDPR being rolled out and increasing media coverage.

    It all depends. It all depends on the programs, requirements. Ability to receive anonymous data.

    3) Do you see other threats related to social media data mining?

    It depends on what to interpret intellectual data. And for what these data are used. If the data is used for medical purposes to protect a person, or to help the user, then probably yes.

    ReplyDelete
  11. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    Maybe with not 100% accurate but I know that similar psychological diagnosis are make now so why in the future wouldn't they more effective?

    2) Will the access of researchers to social media data be limited?

    The access shouldn't been limited. People should can share every information which they want...other things is that people get too many information about theirself in social media.

    3) Do you see other threats related to social media data mining?

    Many but the most is to affect on people when we know what they want to hear and what they believe.

    ReplyDelete
  12. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    Of course not, what people publish in social media is their choice. It is not possible to create a complete psychological portrait on the basis of this.
    2) Will the access of researchers to social media data be limited?

    This depends mainly on the purpose for which the data will be used, but also on the goodwill of the companies. They may want to restrict access to the data in order to use it themselves.

    3) Do you see other threats related to social media data mining?

    I think that in most cases the analysis of these data will not serve a good purpose. Companies will want to discover the weaknesses of people and on their basis influence their behaviour and choices.

    ReplyDelete
  13. Piotr,
    I think it will be possible, but I would never trust it. Good psychological diagnosis give conclusions in a very sensitive and important for humans area. As for me, I would be totally against making life decisions on basis such non-human analysis.
    I don’t think it is definitely possible. Researches will find possibility to gain such data, anyway. All the limitations will be bypassed. I am not optimistic in this respect.
    For me, the most important threat in this respect is that such information is not 100% reliable and trustworthy. This is not real and total honest data. This is the data prepared and selected. That’s the reason, why I couldn’t fully trust information flowing from this source, and that’s why I cannot trust any analysis based on this kind of data.
    With kind regards,
    Marta

    ReplyDelete
  14. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I think that even real life psychological diagnosis are not 100% accurate. Human's psyche is too complecated to diagnose it just throug social media. Of course some clues or preliminary diagnosis may be based on social media but it is never enough to make full diagnosis.
    2) Will the access of researchers to social media data be limited?
    I think it depends on the social media privacy policy or the governments regulacies. In my opinion it should not be because the work of researchers shouldn't be limited so the people knowledge may grow.

    3) Do you see other threats related to social media data mining?
    I think there are many threats like social media engineering an manipulating of people views. Everything depends of how much we share.

    ReplyDelete
  15. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I don’t think so. However I suppose it may have high accuracy (maybe about 90-95%).

    2) Will the access of researchers to social media data be limited?
    Yes, to my mind it is limited now and will be limited in the future. Only the limit will be moved. In my opinion everything is limited (even the infinity is potentially limited – as it is called by logic).

    3) Do you see other threats related to social media data mining?
    I see some threats related to this topic but all of them have already been mentioned. First of all phishing attacks, afterwards selling our privacy and finally influencing our subconscious and our decisions.

    ReplyDelete
  16. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    If people will share 100% accurate information about themselves...

    2) Will the access of researchers to social media data be limited?
    As we all know, Facebook had a problem with that recently, so it's hard to say what will happen in a near future.

    3) Do you see other threats related to social media data mining?
    Losing privacy completely.

    ReplyDelete
  17. I doubt it will ever be possible to have 100% psychological diagnosis even based on one-on-one sessions, but inspired by the article I can think of some general profiles or psychological patterns that should be possible to asses with a high accuracy.

    The access is already limited, for example Facebook changed its API after the last scandal.

    I think that aggregated data aren't too much harmful. It can be used to analyse customer base, create better products etc. What can be harmful is collecting information about individuals as it can be used in social engeeniring attacks or other criminal activities.

    ReplyDelete
  18. Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I don't think it will be accurate because people don't always tell truth in social media especially. Such diagnosis may indicate some serious evidence but I don't think we should believe it.

    Will the access of researchers to social media data be limited?
    It should be limited. However, some people are ok with the fact that their data is collected and exploited. Nonetheless, aggregated data should not be harmful.
    Do you see other threats related to social media data mining?
    We lose privacy and are minds/opinions are vulnerable to any further attacks thanks to such research. Phishing is also a good example.

    ReplyDelete
  19. 1. Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    I am not an expert in this field, but I think that we will not be able to give such a diagnosis. The index will probably increase from 88%. Science is moving forward and new methods and tools are constantly emerging.

    2. Will the access of researchers to social media data be limited?
    I think not. Data are useful for researchers to determine human preferences and develop new technologies to make life easier.

    3. Do you see other threats related to social media data mining?
    I don't see any other threats. By accepting the regulations of social networks we accept the processing of our data.

    ReplyDelete
  20. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    Hi, it is completely unknown field for me. In my opinion even physical presence in a hospital do not give us 100% accurate diagnosis. Nevertheless we can imagine that the psychological diagnosis based on our social media activities might be more accurate. BUT … are you sure that the person who is active social media user writes that what really thinks? A lot of people pretend to be someone different (better?) in virtual reality.

    2) Will the access of researchers to social media data be limited?
    I think that it will not be. I think that the access to social media data should be open.

    3) Do you see other threats related to social media data mining?’
    It gives a possibilities for criminals that can collect informations about us and use it. FOr instance people loves live streams - and gives information: hello my dear crim … we are currently on holidays, take a look on our goods that we left at home.

    The next point - if some kind of software classify us as an idiot - and this information will be widespread by accident … I would not be happy. ;)

    ReplyDelete
  21. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    No. Psychology is sadly not doing well, as consciousness evolves independently of any measurments written or described by certain group of people so it is inadequate due to delay. Most of social media accouts are inadequate because they are a selective database of a person's life -> false ego that a person wants to show to the world. So how are you supposed to learn about a person having a small percentage of knowledge about him/her?
    2) Will the access of researchers to social media data be limited?
    I dont know, BUT I'd say it will increase, because reaserchers' need to know, to have a sense of control is a constant.
    3) Do you see other threats related to social media data mining?
    Tons. Social media are a bit like going to a bazaar to a guy that says that you can put your photos, information, privacy into his pocket and it will be safe. :))))))) ... Think of it a bit.

    ReplyDelete
  22. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    It is hard to make diagnosis based any kind of data, we can be close to 100% but I do not believe in 100% accuracy. There is to many features that effects possible outcomes.
    2) Will the access of researchers to social media data be limited?
    yes, even General Data Protection Regulation take care of this
    3) Do you see other threats related to social media data mining?
    As far as I know social media crowling is illegal and for sure most of the largest social media operators do not allow this procedure in term of use

    ReplyDelete
  23. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    As long as we have people like me, it won't be possible. Let's avoid posting too much data about ourselves and it will never be accurate. In my honest opinion, professionals can't be 100% accurate, so how algorithms can be?
    2) Will the access of researchers to social media data be limited?
    Maybe, who knows? You can scramble data and still get a lot of good results. This doesn't mean it will be accurate. Of course, we can expand our datasets and get data from real inquiries, but it's not as free as data in social media.
    3) Do you see other threats related to social media data mining?
    Our privacy is compromised. Even though, we are accepting terms and so on. That's why uploading data about us shouldn't be so common. At least from my perspective.

    ReplyDelete
  24. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    I really hope it won't be possible. I do hope that people will come to their senses and stop publishing all the private stuff and start thinking twice before posting and commenting, and putting photos online.

    2) Will the access of researchers to social media data be limited?

    Even if it will be limited, still they would gather enough data to conduct other weird researches and diagnosis.

    3) Do you see other threats related to social media data mining?

    Compromising one's privacy is a sufficient threat to me. One does not need other threats to be enumerated, this is the most serious one, as it triggers all kinds od dangerous consequences.

    ReplyDelete
  25. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?
    In future? In my opinion we can make it already now. There is so much people that can not breath without putting this information on Facebook. Facebook contains every information about their lives. I’m sure that we can make psychological diagnosis based on that.

    2) Will the access of researchers to social media data be limited?
    It should be limited. It contains so much personal data. My dream is to disappear from every social media. I want to be a ghost.

    3) Do you see other threats related to social media data mining?
    There is so much threats that it is hard to count them all.

    ReplyDelete
  26. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    I am sure it won’t be possible. On the social media people tend to share only a narrow part of their life, mostly happy events or the most exciting. Their main goal is to show how their life is full of excitements and interesting things, so it can’t be a proper base to define some psychological diagnosis, which can be seen as accurate and can be the ground to prescribe proper meds.
    2) Will the access of researchers to social media data be limited?

    I think the access will be limited, because first of all different researchers may have different goals in their mind in case of using such data and only part of them can be seen as proper and important for the humanity, so only some of them should have access to such data. The other reason is that the data will be offered through the Internet, so there won’t be enough verification of the researchers if everyone could just say that he is a researcher and want some data. So I am sure that social media data will be limited.
    3) Do you see other threats related to social media data mining?

    Of course I see other threats. There is just so many of them – the possibility of selling them to the big companies which then use it for their own purposes, mainly to sell their products in more effective ways, threat of using the photographs gathering by the social media platforms by the pedophiles or hackers, using this data by the thieves to schedule the next robbery… There is just a fraction of possible threats.


    ReplyDelete
  27. 1) Will it be possible in the future to make a psychological diagnosis based on social media that will be 100% accurate?

    I think for some disorders, like social media addiction, yes. A guaranteed 100% accurate test in psychology doesn't sound real to me. There will be of course a lot of hype to be expected -- and many people will be taking action based on such diagnoses.

    2) Will the access of researchers to social media data be limited?

    Yes and no. Right now there are a few APIs or datasets which are available, but still it is what the social media platforms want to be made accessible. So, indepentent researchers may have a limited access to such data. But what happens inside the companies is a completely different story. There is of course a lot of regulations, especially in the EU, but also in the US to give some protection to users, but I am worried that they are not keeping pace with the R&D departments of said companies, and instead stifle possible competition. And a diversification of the market would go a long way toward a healthier environment.

    3) Do you see other threats related to social media data mining?

    The data trail lives on much longer than we realize. There may be a situation in the future where innocent people could be harmed.

    ReplyDelete