I prefer R, sometimes I use Eviews.

Yes, i agree to the point that now a days R and Python softwares are in huge demand. Technological advancement in econometric softwares are taking place due to availability of High Frequency data and to provide more robust forecasts.

i have also used Eviews and SPSS for data analysis, but right now i am using R software for ultra high frequency data filtration, believe me, the codes which we use in R software are very useful to perform any required data sets as per our objectives. There are plenty of packages and commands to help and perform the data filtration and analysis.

R in Action is an excellent book (User Manual) to get a clear idea about R software.

Thanks @ Adriano for initiating this discussion, it is really helpful.

• Although many softwares applications are limited to types of data we used, I found that STATA is modern and general command driven package for statistical analyses, data management and graphics. For example, if you take Eviews it is common for Time series and Panel analysis; SPSS limited to explanatory analysis (descriptive analysis); Gertl limited to time series and panel data. However, STATA is INCLUSIVE, i.e, it provides commands to analyze data (cross-sectional, time-series, longitudinal, repeated-measures, and correlated data), survival-time data, cohort study, and the like. Furthermore,
• It is used by many field of studies like economists, social scientists, medical researchers, biostatisticians, epidemiologists, sociologists, political scientists, geographers, psychologists,, and other research professionals needing to analyze data.

Thus, my conclusion is that the software you use is based on types of your data.

EViews, RATS.

Today the main econometric/statistical packages that I use are R, Gretl, Octave and Matlab. I have also used RATS, Stata, Eviews, Mathematica and several other packages.

During my working days one of my duties was to provide advice to economic colleagues on econometric/statistical analysis and software for such analyses, I was often asked this question by various colleagues. My answer depended on

1. The topic of the study and the dept of the required analysis

2. The colleague's knowledge of econometrics.

3. The colleague's knowledge of computing.

My answer would also take into account that I might be required to help and I would then recommend a package that I was familiar with. If ther are people in your organisation that are familiar with a particular piece of software use that and avail of whatever help you can get.

If the colleague was not very experienced with econometrics or computing I used to recommend microfit but would today prefer Gretl, Gretl is free, easy to use and well documented.

RATS used to be very good for time series analysis and Stata for microeconometrics. Rats now has good coverage of microeconometrics and Stata of time series. Being more familiar with thier original uses I would tend to pick RATS for time series analysis and STATA for microeconometrics. I do not recommend Eviews as I think replication is important and replication is difficult in Eviews. If you use the GUI in gretl, rats or stata it is possible to record the menue items as commands and rerun the analysis with possible amendments. With Eviews this is not possible and you must separately write the instructions using the Eviews command language. However Eviews is a MS windows only package and some some windows users find the GUI easy to use.

R is probably the most general package available but it is difficult to use. If you want to do advanced work in econometrics you should also look at programs such as matlab, gauss, octave. Julia or Scilab. I have used python from time to time but have not had the time to use it in depth. It is reported as being a suitable alternative to R.

To summarise I would start with gretl and then move to R. Once you have learned one lanuage it is relatively easy to learn a second

There is no best econometric software. However, R might be one of the best options for econometricians (and widely used among academics worldwide), although, it takes a while to get familiar with.

From my personal experience, MLwiN is very efficient for the Bayesian estimations of complex non-linear multilevel models (e.g. those including cross-classifications), however, it is not commonly used by economists - more by social scientists.

There is nothing like best Economic Software because the best use of software is depend on data you are working with. If data is time series and econometrics analysis are required then Eviews is preferable and if you are familiar with command writing skills then use of R software packages always recommended.

Good question, however, I tend to agree more with Vasily, "there is no best econometric software" but having flexibility on two or more softwares keeps you going. As SPSS is efficient for descriptive statistics, STATA may be good for household survey and R or Gretl might help in complex module building and data visualization.

I think Gretl for Time Series and Statsoft for regression analysis.

if I am using panel data and have to analyse the effects of moderators and mediators in my regression analysis. which software will be the most appropriate?

Ronoowah Rishi Kapoor , for panel data, I would say that you have some good options: R::plm, Eviews, and Stata are may best tips.

Renato Rossetti , Gretl is very interesting. I really don't understand why it has little highlights in economics compared to other like R or even those not free like Eviews and Stata.

thanks for your tips. i was trying to use SPSS for panel data analysis but does not appear to be that helpful. I think i will opt for eviews which seems to be more user friendly. Can you please help by advising on any textbooks or others on how to use eviews to analyse mediator and moderator effects for panel data anlysis.

is Gretl also good or better the eviews to analyse mediator and moderator effects in regressions for panel data?

Ronoowah Rishi Kapoor , have you seen this post: https://www.researchgate.net/post/How_mediation_analysis_can_done_using_STATA ?

and also

https://stats.idre.ucla.edu/stata/faq/how-can-i-do-moderated-mediation-in-stata/

Thanks a lot for your guidance. I will also consider gretl which appears to be as equally good as eviews and available for free.

I think R still the best, its great project

Very interesting question professor Adriano Marcos Rodrigues Figueiredo. When considering this question we could also think in some interesting points about why R and Python are becoming so popular in Econometrics and Statistics:

- Machine Learning - The name can cause an impact in unadvertised people, but an important part of the predictive models for machines are based in linear regressions, time series, panel data, least squares and other things that composes the econometrician's toolbox. R and Python are both widely used in Machine Learning because of their versatility in handling statistical analysis and algorithms, data cleansing and some other big data applications (most done in Python-based libraries).

- Big Data - Although many software can handle big datasets (millions of rows and hundreds of columns), few of them can do this with the agility and accuracy of Python and R, but remaining free of charge. Of course, we should remember that R depends a lot of the available computer memory depending of the application, so other solutions like Teradata, Scala, Julia Computing or other software could be better.

- Data Science - Artificial Intelligence, Machine Learning, Big Data, Hyper Data, Data Mining and many other concepts are becoming more and more known today, and the advances on these areas can contribute a lot with the study of Economics and Social Sciences. In summary, two of the few software that can walk through all the aforementioned topics are R and Python, so that is a possible explanation for the popularity of these languages.

PS.: It is worth to remember that almost all the most relevant and accessed courses in Data Sciences, Big Data, Machine Learning, Artificial Intelligence and similar topics in Coursera have at least one module dedicated to learning R, Python, or them both.

for discrete choice models, LIMDEP is the most suitable choice, especially to find marginal effects of MV models

Well, I understand that each software can do "anything" if you're willing to program a script/routine/.do/.prg file etc.

The point is, to me, future possibilities or opportunities to enhance your computational abilities. Then, I believe Python and R are better. For example, when I'm calculting some Leontief's input output multipliers, economic linkages, global value chain decompositions, or deterministic indicators, seems to me that R and Python are better... But if you have intensive use in some specific kind of model (such as discrete choice, or time series, or panel data, or complex survey data, etc) maybe your entry cost will lead to a software like Jamal Shah said above...

Below I include an interesting link:

https://www.r-bloggers.com/whats-the-best-statistical-software-a-comparison-of-r-python-sas-spss-and-stata/

Customer service is at times as important as functionality and usability.

My experience with EViews has been terrible. I am not saying that the others are necessarily better, but if one of the others perform well in this area, especially with regards to educational institutions, I would stick with that and steer well clear of EViews.