Keywords

1 Introduction

Study of the network is the field of the discrete mathematics; we called as the graph theory which is originated by the mathematician Euler. Euler published their paper on the topic of Konigsberg bridge problem in 1736 [1]. Some examples of networks are as follows: road network, railway network, computer network, social network, Internet network, gene network, neural network, and biological network. Social network comes under the complex network, having some properties such as random graph, small-world, and scale-free networks. Paul Erdos and Alfred Renyi gave the theory of random graph in 1959 to understand the properties of the graph, how growing the number of nodes in the given graph. Ducan Watts and Steven Strogatz gave the model for the small-world network in 1998. The important property of the small-world network is high clustering coefficient [2]. The edge implementation with the probability of P, when p = 0 network work such as regular lattice and when p = 1 network work as random graph. Barabasi and Albert first introduced the scale-free networks which follow the power law degree distribution. Scale-free network divides into two parts, static scale-free network and evolving scale-free network.

Social network is a type of complex network. A social network is a set of people or groups of people having interaction among them. There are so many social network applications, i.e., Facebook, MySpace, LinkedIn, Flicker, Twitter, and YouTube. To understand dynamics of the user activities, they should know the mechanism of interaction among the online social network users. Online social networks are basically designed for two primary purposes; these are sharing and interaction of data and support the social activities of users. The main aim of social network analysis research is to understand the dynamics of network and its structural properties. Facebook is one of the most popular social networks [10]. There are 1.44 billion monthly active users till March 2015 at the Facebook. Facebook profile has so many attributes such as birth date, hometown, contact information, college, employers, high school. The average Facebook user makes more than four attributes set as publicly. Online social networks are investigated to find the relationship between modeling phenomena and characteristics of real networks, i.e., average path length, clustering coefficient. One common question is how to connect the local structure on phenomena with the global dynamics. Visualization methods and tools are used to analyze evidence parameter of nodes. The position of nodes is determined by the graph layout method to remove these issues, but it becomes more complicated when number of nodes increases in the networks. To address this issue of large number of nodes, community structure of the network is used. Community structure connects the local network to the global.   Social networks come under assortative, and biological and technological network comes under disassortative [15].

In this paper, the relation between time and Facebook stories is found out (Posts, Comments, and Shares). Proposed work is based on the review literature and analysis of Facebook networks. The rest of this paper is organized as follows. The related work is summarized in Sect. 2, analysis and experimental results are covered in Sect. 3, and finally, Sect. 4 concludes the paper.

2 Related Work

Salamanos et al. [22] investigated relation between Likes and Communities in social network (Facebook). They used crawler for collecting data to their analysis which is based on breadth-first search and designed in Python. The result is based on two steps, first one is to detect relation between communities and likes and second one is validation with the help of communities detection algorithm. Communities structure referred as partition among the users’ intention of the research is found out that same type of communities performed similar types of activities on social network. They used Louvain algorithm. They used Gephi, social network analysis tool, for all experiments (visualization). Quinn et al. [20] analyzed the behavior of users at social network in perspective of age. They categorized persons into two categories: young user and old users, persons those having 15–30 years old come under young category, and those are 50+ years old come under old users. They found that younger user having 11 times more friends as compared to the old users, and they also considered following terms in their analysis: user comments, user replies, wall comments, status comments, and media application in the perspective of age. They collect data from 250 individual different profiles of Facebook. Functionality of the online users such as reply, post, share, and comment varies from age to age (old users and young users).

Hirsch and Sunder [8] discussed the effects of sharing the stories on social networking sites (Facebook). A total of 70% news comes from friends and family on social networking sites. A new feature of hashtags is available on social network by which anyone can easily following a topic. Basically, there are three types of broadcasting of stories: user can shared a story at own wall, user can shared a story on friends’ wall which is visible to all mutual friends, and user can directly message to friend in this way story is secret between you and your friend.

Nguyen and Tran [19] presented the paper on the Facebook activities, in which they analyze the users’ connectivity, similarity, and activity. They analyzed a user having many contacts is participate in many activities or less activities and vice a versa, activity distribution role in users communities which user takes more participation and same type of activities performed by the similar types of person or not. They used R programming for analysis, for plotting the result ggplot2 and power fitting algorithm in their study of the activity correlation. They collect the data from Max-Planck Software Institute for their research work; data set has two parts: first part having the information of friendship links and second part having the information of wall posting. Eftekhar et al. [3] used an online survey for their research work, in which they ask some question and participants can give answers with the specify rating. They analyzed the photographs related to activity and find out the personality regarding the photographs. Basically, the user performed activities on social networking sites by which he/she described himself/herself behavior. And authors notify this type information and predict the behavior of social network user. In their survey, 115 participants take role actively and put their view of the questions those are in the survey questionnaires. According to their report, 219 million photographs are uploaded daily on average. They analyzed two types of activity in their research: First is user created own profile and upload photographs, videos, and album; second is communication among users’ like, share, and comment. They provide the result of their research at any photographs/videos with the help of five parameters: communication on that particular photographs, visual presence of photographs, extraversion of Facebook, conscientiousness, and openness of the Facebook. Total participation is 130, but 15 out of which not gave response of the survey. So only 115 participants’ reviews only consider in their research work.

Jiang et al. [11] introduced an algorithm for the social information recommendation system with the help of probabilistic factorization matrix. They designed algorithm for two social networks: Facebook and Twitter, where Facebook is bidirectional, but Twitter is a unidirectional social network. They analyzed their result on two data sets: First one is collected from Renren, a social network in China such as Facebook, and second data set is collected from Weibo, a social network in China like Twitter. They designed a novel recommendation system with the help of well-known recommendation system content-based filtering and collaborative-based filtering. Both recommendation systems have own advantage and disadvantage; they implement advantage of both systems in their matrix factorization recommendation system. Content-based filtering is based on the ranking system while collaborative-based filtering is based on the memory-based models. Khadangi et al. [12] measured relation between user’s activities and their profile information available at online social network. They conducted an online survey for their research and strength measured into four levels. They used two models in the analysis of collection of their data set and multilayer perceptron decision tree model. To check the validity or accuracy, they used 10-cross-fold method for validation purpose. They also compared the results generated from multilayer perceptron, with support vector machine to gain more accuracy in their result. Average number of friends in Facebook social network is 130 while in Renren social network having average number of friends is near about 100. There are 37% of Facebook users having 100+ friends. And the maximum number of friends is bounded by 5000 in Facebook social network, but Twitter user can have more than million followers. They removed outlier with the help of local density method.

3 Experimental Result

Here, we used Digital footprints tool for our results and analyzed those considering in our research work. A digital footprint is an online social network analysis tool. Users required taking the permission from Digital footprints for accessing it. Our research work is based on the following Facebook resources (Table 1).

Table 1 Facebook resources [23]
Full size table

Figure 1 shows the Facebook Posts over the experiment which includes Posts on wall, Posts in page, and Posts on groups. Figure 2 shows the Facebook comments over the collected data, and Fig. 3 shows the Facebook shares over the analysis for the results.

Fig. 1
figure 1

Facebook posts

Full size image
Fig. 2
figure 2

Facebook comments

Full size image
Fig. 3
figure 3

Facebook shares

Full size image

Figure 4 shows all the feeds (wall feeds, group feeds, and page feeds) which include 4 participants, 2 groups, 1 page, and duration is one year. All feeds include here are available at wall of participants, group, and page except the Facebook-generated stories. Data are collected from 1–1–15 to 12–1–16.

Fig. 4
figure 4

Number of Post and Comment

Full size image

Figure 5 shows the Facebook post with respect to time (day), which shows average number of post per day entire one year posted by 4 participants, 2 groups, and 1 page. We find that minimum number of post uploaded at Facebook is 4 (2 A.M. and 4 A.M.) and maximum number of post uploaded at Facebook is 58 (12’o clock).

Fig. 5
figure 5

Number of Facebook Stories in a day

Full size image

4 Conclusion

Social networks are rapidly grown in modern era whereas the Facebook is one of the most popular social network all over the world. All dynamics of the Facebook are randomly taken place, so find out a relation among them is a crucial task. It is necessary to collect the data from the Facebook for giving some results regarding social network (Facebook). Collection of data from the Facebook is a big issue because of the privacy concern interrelated to the peoples activities. We used the Digital footprints for the collection of the data from the Facebook. We find a relation between Facebook stories with respect to time, where most of the Facebook stories comes at 12’o clock according to our survey and minimum stories in between 2 A.M. and 4 A.M. Constraint of my work is limited data collection; in the future, the same task can be composed to a large data of amount. When the duration of data collection increased and number of participants, Facebook pages, Facebook groups more over added to the research work that would be more crucial task.