Exploring the Characteristics of Popular Deep Learning GitHub Repositories

This repository is the accompanying repository for our paper "Exploring the Characteristics of Popular Deep Learning GitHub Repositories".

Data

In the data directory, we provide the popular an unpopular Deep Learning (DL) repositories studied in our paper. Furthermore, int the data/model_train directory, we provide the training data for the classifiers in Research Question 3 (RQ3), along with the outcomes obtained from diverse classifiers for two tasks: predicting the popularity status of Deep Learning (DL) repositories and forecasting the cumulative star patterns of DL repositories.

Spearman correlation coefficient between features and popularity

In Research Question 2 (RQ2) of our paper, we conduct an analysis on the correlation between our proposed features and the popularity of DL repositories using Spearman's correlation coefficient. In our paper, we present the Rho values visually through images; however, in this context, we provide the complete results.

Categories	Feature Name	domain	value	type	star	fork	download
Basic information of the repository	num-code-file	all	rho	0.088437	0.144113	0.28033	0.263931
			p-value	0.02561	0.000263	9.1E-13	1.45E-11
		framework	rho	0.760197	0.681541	0.514393	0.427044
			p-value	4.21E-06	9.07E-05	0.00605	0.02631
		software	rho	0.545112	0.308495	0.51992	0.244205
			p-value	2.2E-16	4.15E-09	2.2E-16	4.45E-06
		education	rho	0.180875	0.144725	0.028094	0.02314
			p-value	0.00664	0.03036	0.6758	0.7305
		research paper	rho	0.230696	0.10562	0.318371	0.279988
			p-value	0.1635	0.528	0.05141	0.08865
	num-code-line	all	rho	0.088437	0.144113	0.128002	0.167892
			p-value	0.02561	0.000263	0.00133	2.15E-05
		framework	rho	0.65606	0.681541	0.509494	0.397216
			p-value	0.000203	9.07E-05	0.006637	0.04021
		software	rho	0.282804	0.308495	0.283132	0.145688
			p-value	8.02E-08	4.15E-09	9.2E-08	0.006715
		education	rho	0.249727	0.144725	0.070929	0.05342
			p-value	0.000159	0.03036	0.2905	0.4263
		research paper	rho	0.1635	0.528	0.05141	0.08865
			p-value	0.943	0.528	0.6723	0.9075
	has-discussion	all	rho	0.246571	0.246052	0.159816	0.201223
			p-value	2.82E-10	3.08E-10	5.91E-05	3.24E-07
		framework	rho	0.383886	0.119505	0.119725	0.290932
			p-value	0.04806	0.5527	0.552	0.141
		software	rho	0.292973	0.313826	0.226809	0.208682
			p-value	2.57E-08	2.16E-09	2.17E-05	9.42E-05
		education	rho	0.116838	0.109894	0.104323	0.039466
			p-value	0.08101	0.1009	0.1195	0.5568
		research paper	rho	0.235702	0.226796	0.182991	0.055535
			p-value	0.1543	0.1709	0.2715	0.7405
	number of repositories that they depend on	all	rho	0.245452	0.2689	0.224209	0.230217
			p-value	3.41E-10	5.17E-12	1.42E-08	4.51E-09
		framework	rho	0.417375	0.493033	0.40405	0.395091
			p-value	0.0303	0.008974	0.0366	0.04139
		software	rho	0.372406	0.392834	0.343594	0.205414
			p-value	6.86E-13	2.74E-14	5.75E-11	0.000122
		education	rho	0.048994	0.07054	0.118386	0.016815
			p-value	0.4656	0.2932	0.07704	0.8024
		research paper	rho	0.246528	0.354839	0.305877	0.066532
			p-value	0.1357	0.02882	0.06182	0.6915
	number of repositories that depend on them	all	rho	0.363074	0.361601	0.298788	0.169106
			p-value	2.2E-16	2.2E-16	2.24E-14	1.87E-05
		framework	rho	0.307282	0.424758	0.270141	0.14678
			p-value	0.119	0.02721	0.173	0.465
		software	rho	0.473682	0.483379	0.439504	0.134763
			p-value	2.2E-16	2.2E-16	2.2E-16	0.01223
		education	rho	0.164406	0.160829	0.071575	0.149863
			p-value	0.01375	0.01598	0.2861	0.02489
		research paper	rho	0.292477	0.343494	0.203167	0.068912
			p-value	0.07476	0.03474	0.2212	0.681
	num-topics	all	rho	0.476241	0.489373	0.427254	0.257087
			p-value	2.2E-16	2.2E-16	2.2E-16	4.99E-11
		framework	rho	0.453428	0.484945	0.242764	0.342366
			p-value	0.01753	0.01035	0.2224	0.08045
		software	rho	0.488971	0.486733	0.461323	0.284391
			p-value	2.2E-16	2.2E-16	2.2E-16	7.68E-08
		education	rho	0.510979	0.538203	0.42396	0.135018
			p-value	2.68E-16	2.2E-16	3.48E-11	0.04352
		research paper	rho	0.314684	0.360525	0.258396	0.17271
			p-value	0.05432	0.02618	0.1173	0.2998
	num-contributor	all	rho	0.769704	0.790419	0.714384	0.301072
			p-value	2.2E-16	2.2E-16	2.2E-16	9.46E-15
		framework	rho	0.795577	0.839102	0.594954	0.211531
			p-value	7.05E-07	4.55E-08	0.001063	0.2895
software		rho	0.807626	0.798946	0.695819	0.277669
		p-value	2.2E-16	2.2E-16	2.2E-16	1.59E-07
education		rho	0.716777	0.779027	0.694654	0.104043
		p-value	2.2E-16	2.2E-16	2.2E-16	0.1222
research paper		rho	0.634965	0.704492	0.514619	0.09344
		p-value	0.000216	1.99E-05	0.006024	0.643
Features of DL- related repository	has-framework	all	rho	0.11319	0.144799	0.120469	0.041572
			p-value	0.004232	0.000246	0.002535	0.296
		framework	rho	0.042798	0.038758	0.009707	0.288308
			p-value	0.8321	0.8478	0.9617	0.1447
		software	rho	0.121078	0.152464	0.120115	0.017122
			p-value	0.02389	0.004363	0.0259	0.7513
		education	rho	0.095761	0.122257	0.174061	0.070171
			p-value	0.1532	0.06779	0.009041	0.2957
		research paper	rho	0.358569	0.388832	0.507635	0.140808
			p-value	0.02706	0.01584	0.001141	0.3991
	num-framework	all	rho	0.121675	0.151717	0.133303	0.029837
			p-value	0.002096	0.000121	0.000827	0.4533
		framework	rho	0.081809	0.011662	0.031957	0.255757
			p-value	0.685	0.954	0.8743	0.1979
		software	rho	0.133626	0.165314	0.140478	0.021161
			p-value	0.0126	0.001974	0.009082	0.6953
		education	rho	0.108579	0.131771	0.182769	0.071099
			p-value	0.1051	0.04887	0.006083	0.2894
		research paper	rho	0.283401	0.433548	0.490942	0.02286
			p-value	0.08467	0.006544	0.00175	0.8916
	has-model	all	rho	0.159323	0.153248	0.138749	0.113147
			p-value	5.37E-05	0.000103	0.000499	0.004338
		framework	rho	0.006221	0.078872	0.079017	0.076271
			p-value	0.9754	0.6958	0.6952	0.7053
		software	rho	0.192457	0.183124	0.176054	0.080073
			p-value	0.000305	0.000597	0.001041	0.1377
		education	rho	0.109123	0.106185	0.143548	0.02752
			p-value	0.1033	0.113	0.03175	0.6821
		research paper	rho	0.339683	0.350194	0.29733	0.16007
			p-value	0.03693	0.03113	0.06985	0.3371
	num-model	all	rho	0.162549	0.15965	0.143967	0.129987
			p-value	3.76E-05	5.18E-05	0.000302	0.001036
		framework	rho	0.048636	0.009278	0.150509	0.009981
			p-value	0.8096	0.9634	0.4536	0.9606
		software	rho	0.198669	0.192869	0.182368	0.108837
			p-value	0.000192	0.000296	0.000677	0.04336
		education	rho	0.109852	0.107663	0.144642	0.027517
			p-value	0.101	0.1081	0.03046	0.6821
		research paper	rho	0.353967	0.365596	0.317646	0.158461
			p-value	0.02924	0.024	0.05197	0.342
	has-pretrained	all	rho	0.232614	0.206773	0.219134	0.103144
			p-value	2.83E-09	1.4E-07	3.04E-08	0.009352
		framework	rho	0.309344	0.257382	0.368366	0.08023
			p-value	0.1164	0.1949	0.05868	0.6908
		software	rho	0.280941	0.241997	0.273027	0.086643
			p-value	9.83E-08	4.97E-06	2.7E-07	0.1082
		education	rho	0.115475	0.127106	0.143412	0.019327
			p-value	0.08464	0.05751	0.03191	0.7736
		research paper	rho	0.23355	0.221158	0.199097	0.091713
			p-value	0.1582	0.1821	0.2308	0.5839
	has-data	all	rho	0.207477	0.209978	0.186767	0.052764
			p-value	1.27E-07	8.85E-08	2.54E-06	0.1846
		framework	rho	0.090685	0.277855	0.095988	0.256578
			p-value	0.6528	0.1605	0.6339	0.1964
		software	rho	0.26565	0.230314	0.226546	0.041131
			p-value	0.26565	0.230314	0.226546	0.041131
		education	rho	0.206226	0.261377	0.23107	0.065231
			p-value	0.001918	7.52E-05	0.00049	0.3311
		research paper	rho	0.052632	0.079582	0.144217	0.235616
			p-value	0.7537	0.6348	0.3877	0.1544
Information of README file	has-quickexample	all	rho	0.612073	0.615748	0.549707	0.220574
			p-value	2.2E-16	2.2E-16	2.2E-16	2E-08
		framework	rho	0.525885	0.392202	0.208741	0.328216
			p-value	0.004842	0.04303	0.2961	0.09464
		software	rho	0.598681	0.617114	0.544512	0.230922
			p-value	2.2E-16	2.2E-16	2.2E-16	1.48E-05
		education	rho	0.705385	0.661132	0.575622	0.057268
			p-value	2.2E-16	2.2E-16	2.2E-16	0.3936
		research paper	rho	0.406181	0.449325	0.416044	0.15039
			p-value	0.01139	0.004656	0.009378	0.3675
	has-newbutton	all	rho	0.457398	0.455374	0.402469	0.271174
			p-value	2.2E-16	2.2E-16	2.2E-16	3.78E-12
		framework	rho	0.399012	0.335342	0.287965	0.307894
			p-value	0.03923	0.08728	0.1452	0.1182
		software	rho	0.509235	0.494608	0.461244	0.274413
			p-value	2.2E-16	2.2E-16	2.2E-16	2.24E-07
		education	rho	0.393322	0.409627	0.344156	0.123129
			p-value	1.05E-09	1.79E-10	1.26E-07	0.06584
		research paper	rho	0.475191	0.562991	0.406124	0.111963
			p-value	0.002571	0.000234	0.363206	0.5033
	has-photo	all	rho	0.433736	0.435603	0.365125	0.178641
			p-value	2.2E-16	2.2E-16	2.2E-16	6.02E-06
		framework	rho	0.725476	0.670684	0.441546	0.29649
			p-value	1.85E-05	0.000129	0.02113	0.1332
		software	rho	0.499909	0.471451	0.449031	0.155894
			p-value	2.2E-16	2.2E-16	2.2E-16	0.003698
		education	rho	0.318912	0.381199	0.225927	0.139783
			p-value	1.09E-06	3.69E-09	0.000658	0.03656
		research paper	rho	0.376867	0.355227	0.312257	0.190277
			p-value	0.01969	0.02863	0.05631	0.2525
Features of the repository owner	user-or-organization	all	rho	0.350304	0.371675	0.320915	0.221329
			p-value	2.2E-16	2.2E-16	2.2E-16	1.78E-08
		framework	rho	0.573539	0.550674	0.485667	0.150254
			p-value	0.001762	0.002916	0.01022	0.4544
		software	rho	0.518511	0.496938	0.488485	0.215252
			p-value	2.2E-16	2.2E-16	2.2E-16	5.56E-05
		education	rho	0.079712	0.174905	0.087178	0.191383
			p-value	0.2369	0.009015	0.2029	0.004211
		research paper	rho	0.246432	0.040705	0.030271	0.366395
			p-value	0.1975	0.8339	0.8809	0.06015
	followers	all	rho	0.111353	0.098405	0.128403	0.099344
			p-value	0.005	0.01318	0.001284	0.01233
		framework	rho	0.446739	0.471668	0.405732	0.059168
			p-value	0.01949	0.013	0.03575	0.7694
		software	rho	0.117516	0.095758	0.125792	0.094119
			p-value	0.02861	0.07484	0.0196	0.08086
		education	rho	0.440293	0.408099	0.433008	0.131944
			p-value	6.08E-12	2.55E-10	3.08E-11	0.0496
		research paper	rho	0.563969	0.686189	0.576776	0.190987
			p-value	0.001441	3.97E-05	0.001636	0.34
	num-public-repository	all	rho	0.231702	0.226832	0.215813	0.066457
			p-value	3.57E-09	7.67E-09	4.95E-08	0.09455
		framework	rho	0.312553	0.334965	0.204014	0.480189
			p-value	0.1124	0.08765	0.3074	0.01125
		software	rho	0.28172	0.255307	0.269518	0.00035
			p-value	9.43E-08	1.44E-06	3.88E-07	0.9948
		education	rho	0.11503	0.163786	0.205319	0.026382
			p-value	0.08728	0.01456	0.002484	0.6958
		research paper	rho	0.416469	0.28898	0.046281	0.157315
			p-value	0.02462	0.1284	0.8187	0.4332