Extracting communities from big data. An Instagram sharing hashtag selected group (exploration phase)

Data collection was aimed at extracting social communities through a visual algorithm (Fry, 2008), and involved the following steps: 1.Acquire - generally concerns how the user downloads data and obtains the data in the first place. 2.Parse - in the current phase, we have 3,000 occurrences to be parsed (counted) and converted into a format that tags each part of the data with its intended use. The final number in the new format is 553. 3.Filter - this step involves filtering the data to remove portions not relevant for our purposes. The final database includes 384 records. 4.Mine - this refers to data mining. The data receive a simple treatment: the pro-gram must figure out the minimum and maximum values for latitude and longi-tude by running through the data so that it can be presented on a screen at an ap-propriate scale. In our research this involves sub-dividing the 384 records into the two groups of the fondazioneplart exploration (248) and exploitation (140) alters. Each category corresponds to a specific semantic class. In addition, we ranked the number of sharing for each semantic class. 5.Represent - this step determines the basic form of a set of data. Some data sets are shown as lists, others are structured like trees, etc. For our data, we chose dif-ferently shaded bargraphs for each dataset. The darker or lighter shading and the length of each bar correspond to the sharing and frequency of hashtags. Data elaboration and visualization were achieved using the open source software, Gephi.