4 - Challenges

600 PARTICIPANTS

The First and the biggest data science hackathon in israel

???

Identify the owner of an Instagram account only from his images.

The data

{
    "tlrb": [
        184,
        252,
        315,
        246
    ],
    "loc": "",
    "tags": [
        "instadaily"
    ],
    "img_id": 0,
    "f_id": 0,
    "rec_id": 0,
    "hw": [
        800,
        640
    ],
    "likes": 508,
    "u_id": 0,
    "sig":[-0.05958271,0.14901283,0.11618581,-0.06251568,-0.17507978,-0.03812096,-0.0193905,-0.11224487,0.146091,-0.16002028,0.16618837,-0.03008917,-0.17222928,0.11777229,-0.09234431,0.18035942,-0.14862281,-0.12064842,-0.17688282,-0.11498997,0.01787003,0.01083076,0.03875625,0.07587616,-0.15488936,-0.29144832,0.00793489,0.0066907,0.10672068,-0.04947066,-0.02109046,0.08000648,-0.18165341,0.0230958,0.02830132,0.07302031,-0.11127843,-0.09616276,0.2402668,-0.01025365,-0.25766936,0.00471468,0.11721811,0.24272521,0.21173681,0.01332821,0.07637604,-0.11739585,0.10150138,-0.34677315,0.06375571,0.15538976,0.02472468,0.09722564,0.07808027,-0.20165823,0.01993477,0.07838429,-0.15724725,0.01968169,0.07694335,-0.04277663,-0.09129029,-0.09475227,0.24360304,0.17448178,-0.12813224,-0.14937787,0.19624516,-0.23212805,-0.05501983,0.0706668,-0.03906614,-0.13846312,-0.2616255,-0.04263511,0.40869445,0.2237128,-0.15411958,0.03473435,-0.07894515,-0.03232813,0.03911146,0.13297695,-0.10399361,0.02904507,-0.03878252,0.06380285,0.20034961,-0.04289085,-0.03510578,0.25192747,0.12663203,-0.06508642,-0.01034052,0.06629144,-0.21102458,-0.05120991,-0.19535798,-0.07796262,0.03477798,0.02852866,-0.00332176,0.14494044,-0.24536325,0.27188173,-0.06984118,-0.05489121,0.01852174,0.01815532,-0.05828224,-0.06235511,0.12545861,-0.20510219,0.11102055,0.19054906,0.04102717,0.13733476,0.12089179,0.04591015,0.0902166,0.01710343,-0.13160065,-0.09849624,0.02390091,-0.12860009,0.01759114,0.05448996]
}

For each profile with K faces of the owner:
- Each correctly-labeled face, will grant you 1/K of a point.
- Each incorrectly-labeled face, will cost you 2/K of a point.
The final score is the average of the per-profile scores.