Finding most linear region of curve + best linear fit at that point + error + plotting

Hi there, in the first part of my problem I am looking to find the most linear region of a curve; I tried using the "erf", "polyfit" and "polyval" functions to get a fit for my data but the erf function just gave me a n by 1 matrix of "1's", so I have tried to go the long way round by determining the first and second differentials of my data and find the range over which the minimum rate of change occurs but am having trouble "finding" the most linear section... as you will be able to see from my code I am attempting to use "findstr" to search the 2nd differential for a region that is suitably linear but I cannot get it to search for a region in the 2nd differential which is less than or equal to the "lims" (or limits) matrix. Searching the help has taken me quite some time and not really got me anywhere as I am not really sure what I am looking for, nor how to apply much of what I see to the problem I have.

So, assuming I somehow manage to get the most linear section of my data sorted out, I then need to get the equation of that linear section best fit line and also determine the error between the best fit and the actual data (both values being used in later equations).

Ideally I would then like to plot the actual data with the best fit (and errors?).

Please forgive the horrible code, it has been a long while since I used matlab and I was terrible at it then... computational elegance has had to give way to something that makes sense in my head.

    clear all;
  clc;

%Constants and variables
Cw=4185; %Specific heat capacity water
Cg=3086; %Specific heat capacity glycerol
ratiow2g=100; %ratio of specifics 
C=((Cw/100)*ratiow2g)+((Cg/100)*(100-ratiow2g)); %Combined specifics
V=0.001; %Volume
m=0.003; %Mass
Em=0.0005; %Error in mass
ts=5; %Timing interval (s)

%import data
impdat=importdata('T2.xlsx');
dat=impdat.data;
szd=length(dat);
time=[0:5:((szd-1)*ts)]';
data=[time,dat];

%plot(time,dat);

%First differential of data with respect to time
for i=(1:szd-1);

      ddatdtime(1:i:szd)=((dat(i+1)-dat(i))/ts);

end

%Second differential of data with respect to time
for j=(1:szd-2);

      d2datd2time(1:j:szd)=((ddatdtime(j+1)-ddatdtime(j))/ts);

end

szd2=length(d2datd2time);

lims=[0.001 0.001 0.001 0.001 0.001];

res=findstr(d2datd2time, lims); %The section of the data which has the most linear slope?

ddatlin= %The gradient of the most linear portion of the data

Eddatlin = %The error in gradient/fit line

%SLP calculation
SLP=((C*V/m)*ddatlin);

%SLP error calculation
ESLP=(((C*V/m)*Eddatlin)+((C*V/m^2)*Em));

And an example of my data:

dat =

     23.4000
     23.4000
     23.5000
     23.7000
     24.0000
     24.2000
     24.5000
     24.8000
     25.0000
     25.2000
     25.5000
     25.8000
     26.0000
     26.2000
     26.4000
     26.7000
     26.9000
     27.1000
     27.4000
     27.6000
     27.8000
     28.1000
     28.3000
     28.6000
     28.8000
     29.0000
     29.2000
     29.5000
     29.7000
     29.9000
     30.1000
     30.3000
     30.6000
     30.8000
     31.0000
     31.2000
     31.4000
     31.6000
     31.8000
     32.0000
     32.3000
     32.5000
     32.7000
     32.9000
     33.0000
     33.3000
     33.5000
     33.7000
     33.9000
     34.1000
     34.3000
     34.5000
     34.6000
     34.9000
     35.1000
     35.2000
     35.4000
     35.7000
     35.9000
     36.0000
     36.2000
     36.4000
     36.6000
     36.8000
     36.9000
     37.2000
     37.3000
     37.5000
     37.7000
     37.9000
     38.0000
     38.2000
     38.4000
     38.5000
     38.7000
     38.9000
     39.1000
     39.2000
     39.4000
     39.6000
     39.8000
     39.9000
     40.1000
     40.3000
     40.4000
     40.6000
     40.7000
     40.9000
     41.1000
     41.2000
     41.4000
     41.6000
     41.8000
     41.9000
     42.0000
     42.2000
     42.4000
     42.5000
     42.6000
     42.8000
     43.0000
     43.1000
     43.3000
     43.4000
     43.5000
     43.7000
     43.8000
     44.0000
     44.1000
     44.3000
     44.4000
     44.5000
     44.7000
     44.8000
     45.0000
     45.1000
     45.3000
     45.4000
     45.5000
     45.6000
     45.8000
     45.9000
     46.1000
     46.2000
     46.3000
     46.5000
     46.6000
     46.7000
     46.9000
     47.0000
     47.1000
     47.2000