Six things social data taught us this year

As social data becomes more abundant, accessible, archived and sliceable, academic institutions, trade groups and companies in the space are releasing wave after wave of analysis. Combined with survey after survey on consumer use of social, this flood of new research is mirroring the flood of social data itself, and interesting and important findings can easily get lost in our inboxes and Twitter streams. Here are six fascinating findings you may have missed.

1. Tweet text contains reliable gender clues

Of course it’s usually not difficult to guess a Twitter user’s gender from the name behind their handle, but what happens when only the text of their tweets can be used? The Mitre Corporation used a computer to look for patterns in the text of each user in the study’s tweets, and identified gender successfully in 75.8% of guesses. Predictably, more data equaled higher accuracy—when the computer could analyze tweets, handle, bio, and full name, accuracy shot up to 92%.

2. Facebook users confide in more people, are more trusting

A study by The Pew Research Center found interesting social differences between Facebook users and the rest of the US online population.  Frequent Facebookers have “9% more close, core ties in their overall social network,” and “43% more likely than other internet users and more than three times as likely as non-internet users to feel that most people can be trusted.”

3. Price mentions in UGC increase as Dow and CCI fall

As part of his analysis for The Conversation Index, Social Analyst Alex Barrera established inverse correlations between two key economic indices and use of the word “price” in the text of reviews during the Great Recession. Specifically, Alex’s team found a correlation of -.66 between the Consumer Confidence Index and percent of reviews with price mentions, and a stronger correlation of -.68 between the Dow closing index and price mentions. This means that the way we talk about products and services reflects our economic concern.

4. Twitter packs more purchase-driving punch than Facebook

Surveying over 2,500 online buyers, Kantar Media Compete discovered something that should make brands take another look at a channel many of them have been neglecting—Twitter. Thirty-five percent of those surveyed identified Twitter as a significant purchase driver, compared to 23.5% identifying Facebook as such.

5. YouTube links live longer

Over on the Bitly blog, the concept of half-life was applied to links as “the amount of time at which this link will receive half of the clicks it will ever receive after it’s reached its peak.” After running an analysis of 1,000 top bitly links, the bitly science team found that links originally posted on YouTube have a half-life of 7.4 hours, more than double the half-life of Facebook and Twitter links (3.2 hours and 2.8 hours, respectively).

6. Word choice and spelling betray location of Twitter users

Robert Frost is quoted as saying, “You can be a little ungrammatical if you come from the right part of the country.” Carnegie Mellon Machine Learning researchers looked at 4.5 million words in 380,000 tweets and were able to determine where users lived based on their word choice and regional spellings. Here in the South, people don’t just say “y’all,” they tweet it, and in California, people aren’t “really tired,” but “hella tired.” From the CMU press release:

It might be a mistake to assume that the greater interconnectivity afforded by computer networks and sites such as Twitter will necessarily result in more homogeneity in language. The social circles maintained by social networks such as Twitter often are geographically focused, he noted. Also, many people use the Internet to seek out like-minded people with similar interests, rather than expose themselves to a broader range of ideas and experiences.

There’s no shortage of fascinating social data findings coming out every day. This post could have covered hundreds of them, but I chose to zero in on a few that were particularly surprising and unique. What else deserves to be on this list? Let me know in the comments below!

One Response to “Six things social data taught us this year (so far)”

  1. tempat casper

    Data Informasi Pribadi    :
    Personal Information Data

    Tempat Tanggal Lahir;
    Place Date of Birth

    Jenis Kelamin;


    Alamat Lengkap { yang mudah dihubungi melalui teknologi};
    Full Address {which can easily
    be reached through technology}






    Data ini di buat supaya kita saling komunikasi satu sama lainnya
    terhubung dengan baik dan harus saling mengenal supaya kita
    ada rasa
    berkeluarga. Kita adalah satu keluarga untuk mendapatkan hasil yang
    sama.Kepercayaan yang utama andaikata nanti ada ketidakpuasan dalam
    penghasilan di pendirian perusahaan yang akan kita legalkan berbadan
    kita saling komunikasi dengan
    Insya Allah akan berbuat yang terbaik buat
    kalian para anggota yang namanya akan dibuat legal berbadan hukum
    {sesuai nama kalian yang asli/ Kartu identitas}.

    This data is
    made so that our mutual communications with each other well connected
    and must know each other so we have a sense of family. We are one family
    to get the results that the main sama.Kepercayaan later if there is
    dissatisfaction in income in the establishment of companies that we will
    Get legal legal status, let us communication
    with each other. Insha Allah will do what’s best
    for you the members whose names will be made legally
    incorporated {fit
    your real name / identity card}.

    Rencana ini akan berjalan
    sesuai dengan banyaknya anggota sebanyak 300.000 orang yang harus
    saling mengenal satu sama lain yang akan diatur oleh

    This plan will go according to the
    number of members of as many as 300,000 people who should know each
    other that will be governed by

    Ingat jangan memberikan apa yang dapat merugikan anda jika tidak
    dihubungi oleh akan
    mengkonfermasi anda jika sudah tersusun anggotanya sebanyak 300.000
    orang. Dan
    masing-masing anggota mendapatkan satu saham {surat berharga
    yang dapat di tukarkan menjadi Rupiah}. Oleh karna itu angota
    biaya sebanyak maxsimal Rp.100.000 {seratus ribu rupiah} setara dengan US.15,00. Keuntungan
    hasil dapat diambil lima tahun kemudian, dengan hasil 120% per bulan
    seumur hidup dan dapat diwariskan kepada anak pertama/yang diandalkan
    dikeluarga. Pemberitahuan pasti Insya Allah melalui e-mail @ anggota.

    Remember do not give anything that can harm you if not contacted by will mengkonfermasi you
    if it is composed of its members as many as 300,000 people. And each
    member getting one {stock securities that can be exchanged into Rupiah}.
    That’s because members are charged as much as maxsimal 100,000 {one
    hundred thousand dollars}
    equivalent of $ US.15, 00. Gain can be retrieved five years later, with
    results of 120% per month for life and can be passed on to her first
    child / dikeluarga reliable.Notification must Insha Allah via e-mail @

    Kepercayaan dan keyakinan anda dibutuhkan untuk perubahan bersama.

    Trust and confidence you need to change together.

    Terima kasih atas keterbukaan komunikasi yang saling menguntungkan.
    Pastikan kita harus berkomunikasi apapun yang kalian rasakan. Salam kita
    adalah satu keluarga.

    Thank you for mutual openness of
    communication. Make sure we have to communicate anything that you feel.
    Greetings we are one family.

    Nb: Saya berikan referensi kepada cv.Sendang Arta [Ibu Sri Hartati
    62-0819-3210-7414 atau 62-021-9526650] untuk anda supaya ada hubungan
    rekanan. Mungkin anda akan memesan sesuatu yang berhubungan dengan
    percetakan dan perdagangan umum
    {pengadaan barang apa saja yang legal

    diperlukan oleh anda}.Dijamin harga
    sesuai dengan keinginan saudara.
    Beritahukan namanya dari Bobby Martianus…+-@Rp…

    Nb: I give references to cv.Sendang Arta [Ibu Sri Hartati
    62-0819-3210-7414 or 62-021-9526650] for you so that there is a
    relationship partner. Perhaps you will order something related to
    printing and general trade {procurement anything legally required by
    you}.Guaranteed prices in accordance with the wishes of relatives.
    Tell the name of Bobby Martianus …+-Rp!…

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>