Name | Role | GitHub | |
---|---|---|---|
José Roberto Pérez | Developer/PMO | Roberto919 | robertosysm@gmail.com |
Cecilia Avilés | Developer | cecyar | cecyar@gmail.com |
Sergio Sánchez | Developer | esesancr | esesancr@gmail.com |
Erik Palacios | Reviewer | palmoreck | - |
To analyze Mexican doctors' disposition to use technological tools that could enhance their health services by developing three subprojects in order to approach the goal from different angles.
Try to identify if there is already a relationship between doctor's age and his/her perception about the attractiveness of telemedicine. The analysis includes the following sub-analysis:
Recalling what was seen in class, Linear regression attempts to model the relationship between two variables by fitting a linear equation to observed data.
In our case the explanatory variable is the doctor's age, and the dependent variable is the Negative or Positive perception about the attractiveness of telemedicine
Since we don't expect a regression line that must pass through the origin we opted for the regression model with intercept
$$f(x|\beta) = \beta_0 + \beta_1 x$$ $$(x_i,y_i)\forall i=0,1,\dots,m$$
where $\phi_0(x) = 1$, $\phi_1(x) =x$. And we need to calculate: $\beta_0, \beta_1$.
For the linear regression of the negative perception we performed the following calculations:
$$f(x|\beta) = \beta_0 + \beta_1 x$$
$$p_{np}(x) = 0.00432852x + 0.28488185$$
With our beta's adjusted as $\hat{\beta_0} = 0.28488185$, $\hat{\beta_1} = 0.00432852$.
For the linear regression of the positive perception we performed the following calculations:
$$f(x|\beta) = \beta_0 + \beta_1 x$$
$$p_{np}(x) = -0.0031727x + 0.68710041$$
With our beta's adjusted as $\hat{\beta_0} = 0.68710041$, $\hat{\beta_1} = -0.0031727$.
As we could see in both analyses, we corroborated that a relationship already exists between doctor's age and their willingness to offer telemedicine services and we concluded that as doctors get older, they are less willing to offer telemedicine medical assistance services.
Analyze what are the attributes that doctors value the most in a technological platform that could enhance their services. This should be accomplished by examining the answers given to the question "Seleccione las funcionalidades que considera importantes para que estén presentes en una plataforma digital que lo apoye con la gestión de sus servicios médicos."
This analysis must consider the following two conditions:
Text input had to be analyzed for this subproject, which means working with strings mainly.
For this matter, the libraries Pandas and NLTK were used in order to handle the information in dataframes and work with human language data.
The condition in which a default answer was selected was mainly analyzed with Pandas, and the condition for free text was examined with NLTK.
Our sample size consisted of 765 answers given by 765 doctors.
An 82% selected one of the default answers. The top three default answers chosen were:
A 13% choose not to select a default answer, opting to write their own instead. From this:
Analyze whether the results obtained from two distinct survey responses are independent using a statistical hypothesis test. The analysis should encompass the following tests:
The Chi-square test for independence (or Pearson's chi-square test) was used to determine if there is dependence between two selected sets of data.
This project was developed in Python, mainly using the tools scipy.stats.chi2_contingency and scipy.stats.chisquare tools included in the SciPy library.
For all the tests conducted the hypotheses evaluated were defined as follows:
A comparison between alpha ($\alpha$) and the p-value was used to reject or approve the null hypothesis ($H_0$).
Before the results were obtained, we expected for all the proposed tests to yield a result in which the null hypothesis ($H_0$) is rejected. This is because essentially all the variables tested come from the same survey, thus it is logical to think that the variables' behavior would be consistent.
As expected, for all three tests the p-value obtained is lower than the alpha ($\alpha$), thus indicating a dependence among the variables compared.
End of document