Skip to content

Latest commit

 

History

History
62 lines (59 loc) · 2.99 KB

README.md

File metadata and controls

62 lines (59 loc) · 2.99 KB
Raw

FWD_Analysis_Task:

A data analysis task during my Future Work is Digital Scholarship provided by the Egyptian Ministry of Communications and Information Technology.

1- Dataset Description:

  • This dataset collects information from 100k medical appointments in Brazil and is focused on the question of whether patients show up for their appointment. Several characteristics about the patient are included in each row.
  • The data has 110527 rows and 14 columns.

2- InMind Questions:

  • General Questions related to the existence of:
  1. missing values?
  2. wrong datatypes for columns?
  3. complete duplicates in the data?
  4. outliers in each column?
  • Univariate Questions:
  1. Which Gender is more healthier than the other depending on number of enteries?
  2. Which Age values are valid or there are values like 1000..,etc?
  3. Which disease among Hipertension, Diabetes and Alcoholism is most dominant?
  • Bivariate Questions:
  1. Is the number of PatientId same as AppointmentID or may be more than one appointment for the same patient?
  2. Does Receiving Scholarship has strong effect on not cancelling the appointment or not?
  3. Does the period between AppointmentDay and ScheduleDay has effect on cancelling the appointment?

3- Conclusion:

1. Reached Results:

I went through a 110527-row and 14-column dataset about medical appointments inBrazil to discover which features affects the meeting being Canceled or not.

First, the data must be cleaned and I:

  1. dropped AppointmentID and HandCap Columns.
  2. replaced PatientId values with 0 -> 62298 values.
  3. renamed No-show to Canceled for avoiding confusion.
  4. changed the datatypes for Scholarship, Hipertension, Diabetes, Alcoholism and SMS_received.
  5. replaced wrong values in the Age column.

Then, I came with some insights from the analysis that:

  1. In our sample, Females care with their health more than Males as they reserved moreappointments.
  2. The most dominant disease in our sample is Diabetes.
  3. % of Alcohol additction is, by far, higher in Males than Females.
  4. Receiving Scholarship has strong effect on not cancelling the appointment.
  5. LongerReservationPeriod _period between AppointmentDay and ScheduledDay contributes to those cancelled appointments.
  6. Gender has no notable effect on cancelling appointments.

After that, I concluded the main objective that:

Receiving financial Scholarship & getting shorter ReservationPeriod for the appointment arethe most two features that may tend to lower the number of Canceled appointments

2. Limitations:

  1. Handcap has no documentation in the kaggle description & its name has no translation so it may be important but we cannot understand it.
  2. Discription of data on kaggle shows that SMS_received is a field indicates number of message received but in this data it is only binary field with either 0 or 1 which may reflect misleading understand. (Thanks