If you have a decent voltage source and a good heating element it should work fine.
The equation for heat given off by a 'resistor' (any simple element that resists the flow of current) is just the Power equation P=IV. You can rewrite it as P=V2/R where R is the resistance of your heating element and V is the Voltage across your resistor (not the voltage source). You can find the resistance of your coil of wire by using R = pl/A where p = resistivity of the element (greek charactor rho), l is the length of your resistor (the length of the coil if it were stretched out straight), and the cross-sectional area of the wire (estimate it).
Rewriting,
P= V2*A / p*l The resistivity constant of iron is 1.0×10-7. I don't have a number for steel so I'm going to approximate it as iron. For the benefit of this example, lets call the width of the wire 1mm which gives a radial area of 7.85×10-7 m2. Let's again assume a realistic figure of .25m for the stretched out length of wire. If you use a car battery, your Voltage is ~12V (the internal resistance is negligible).
P = 12V * 7.85×10-7 m2 / (1.0×10-7 * .25m) = 376.8 W. At this power output, it would take about 16 seconds to boil one ounce of water (incorrectly assuming all the thermal power is transferred). I used water because I don't know what temperature you need the tobacco to be and 100 degrees Celsius is easy. It takes about 6000 Joules of energy to boil an ounce of water (use Energy = Power * time in seconds for a close estimation). In any case, I doubt someone using this device will want to use it only every 16 seconds (approximately). so for the estimations I gave, it doesn't look feasible. However, you could mess around with that (perhaps find a better material) to get some numbers you like.
