$(1)$ for this project, i am going with multichoice and mixed logit modeling. I need to collaborate a model based on the survey data and then have a preference. then I need to apply the preference data to the public data. I have doubts about how to go about this. I plan to collect survey data on demographics, mode choice preference before and after hike, convenience, income, etc., I am unclear on what data can I apply the preference data on a public dataset. should i use broader demographic data from public dataset to compare individual preference data? how can i get comfort, convenience factors etc. for public data to compare with my survey data. pls guide me$.$

$(2)$ I have done some time series analysis and descriptive analysis, but found the metro ridership has increased after hikes and not much significant changes in other modes. I am stuck and do not know where to focus my further analysis on that will help me study modal shift behavior in this scenario. Here are the results:

Subway Bus

Bike Taxi

$8\mathrm{.}22536-5\mathrm{.}628627$

$16\mathrm{.}29588-8\mathrm{.}49553$

ARIMA for Subway

ARIMA model out$-$of$-$sample MAE $1087088\mathrm{.}0389454397$

trend model out$-$of$-$sample MAE $680571\mathrm{.}5560935534$

ARIMA for Bus:

ARIMA model out$-$of$-$sample MAE $259050\mathrm{.}47807261505$

trend model out$-$of$-$sample MAE $2285879\mathrm{.}3158873995$

My data: I gathered $($daily$)$ ridership numbers data for various modes for $2$ years. I added a column $0$ for non$-$hike period and $1$ for hike period.

Date Subway Bus

Bike Taxi

$2022-12-311927101$

$6864343289560865$

$2022-12-3030634809268816455068970$

The percentage change for hike period $($aug to dec$)2022$ and $2023$ are as follows:

pct$\_$change$()$

$2022$:

Subway $713\mathrm{.}127030$

Bus $1001\mathrm{.}887641$

Vehicle $5\mathrm{.}684320$

Bike $570\mathrm{.}939549$

Taxi $172684\mathrm{.}464693$

$2023$:

Subway $662\mathrm{.}946469$

Bus $908\mathrm{.}874382$

Vehicle $-1\mathrm{.}192719$

Bike $779\mathrm{.}519508$

Taxi $91762\mathrm{.}765317$

and for entire $2$ years of $2022$ and $2023$ are

a

Subway $3676\mathrm{.}343558$

Bus $5121\mathrm{.}723304$

Bike $3984\mathrm{.}841244$

Taxi $264798\mathrm{.}661640$

dtype: float$64$

$(3)$ Please help me how to go about ML regression analysis, what features I should use which will help me analyze modal shift behaviour due to fare hikes. what other ml analysis might help? pls help me with the implementation steps.